diff --git a/.devcontainer-scripts/utils.sh b/.devcontainer-scripts/utils.sh
index 02b588ae..98ac063c 100644
--- a/.devcontainer-scripts/utils.sh
+++ b/.devcontainer-scripts/utils.sh
@@ -32,18 +32,22 @@ config_remote() {
 }
 
 # Setup special .ssh files
-#
+# Prints out lines of text to make things pretty
 # Param 1: bash array, filenames relative to the customization directory that should be copied to ~/.ssh
 setup_ssh() {
+    echo "starting ~/.ssh directory setup..."
+    mkdir -p "${HOME}.ssh"
+    chmod 0700 "${HOME}/.ssh"
+    echo "-----"
     local files=("$@")
-    for file in "${files[@]}"; then
+    for file in "${files[@]}" ; do
         local cfile="/devcontainer-customization/${file}"
-        local hfile="~/.ssh/${file}"
+        local hfile="${HOME}/.ssh/${file}"
         if [ ! -f "${hfile}" ]; then
-            echo "copying ${file}"
+            echo "copying \"${file}\""
             cp "${cfile}" "${hfile}"
             chmod 600 "${hfile}"
         fi
     done
-    ls ~/.ssh
+    echo "~/.ssh directory setup complete!"
 }
diff --git a/.github/workflows/bump_deps.yaml b/.github/workflows/bump_deps.yaml
index 68cb81cb..c94a134d 100644
--- a/.github/workflows/bump_deps.yaml
+++ b/.github/workflows/bump_deps.yaml
@@ -56,7 +56,7 @@ jobs:
           rm -rfv ${{ matrix.variable }}_message.txt
           rm -rfv ${{ matrix.variable }}_commit.txt
       - name: Create Pull Request
-        uses: peter-evans/create-pull-request@v6
+        uses: peter-evans/create-pull-request@v7
         with:
           token: ${{ secrets.UPDATE_BOT_TOKEN }}
           push-to-fork: ci-forks/LocalAI
diff --git a/.github/workflows/bump_docs.yaml b/.github/workflows/bump_docs.yaml
index 218dcc61..6eb390df 100644
--- a/.github/workflows/bump_docs.yaml
+++ b/.github/workflows/bump_docs.yaml
@@ -17,7 +17,7 @@ jobs:
         run: |
           bash .github/bump_docs.sh ${{ matrix.repository }}
       - name: Create Pull Request
-        uses: peter-evans/create-pull-request@v6
+        uses: peter-evans/create-pull-request@v7
         with:
           token: ${{ secrets.UPDATE_BOT_TOKEN }}
           push-to-fork: ci-forks/LocalAI
diff --git a/.github/workflows/checksum_checker.yaml b/.github/workflows/checksum_checker.yaml
index 4f95a4e2..7b85ad35 100644
--- a/.github/workflows/checksum_checker.yaml
+++ b/.github/workflows/checksum_checker.yaml
@@ -36,7 +36,7 @@ jobs:
           sudo chmod 777 /hf_cache
           bash .github/checksum_checker.sh gallery/index.yaml
       - name: Create Pull Request
-        uses: peter-evans/create-pull-request@v6
+        uses: peter-evans/create-pull-request@v7
         with:
           token: ${{ secrets.UPDATE_BOT_TOKEN }}
           push-to-fork: ci-forks/LocalAI
diff --git a/.github/workflows/release.yaml b/.github/workflows/release.yaml
index 5c883db4..a1318b19 100644
--- a/.github/workflows/release.yaml
+++ b/.github/workflows/release.yaml
@@ -294,7 +294,7 @@ jobs:
           export C_INCLUDE_PATH=/usr/local/include
           export CPLUS_INCLUDE_PATH=/usr/local/include
           export PATH=$PATH:$GOPATH/bin
-
+          export SKIP_GRPC_BACKEND=backend-assets/grpc/whisper
           make dist
       - uses: actions/upload-artifact@v4
         with:
@@ -327,7 +327,7 @@ jobs:
           cache: false
       - name: Dependencies
         run: |
-          brew install protobuf grpc
+          brew install protobuf grpc libomp llvm
           go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@1958fcbe2ca8bd93af633f11e97d44e567e945af
           go install google.golang.org/protobuf/cmd/protoc-gen-go@v1.34.2
       - name: Build
@@ -336,7 +336,7 @@ jobs:
           export C_INCLUDE_PATH=/usr/local/include
           export CPLUS_INCLUDE_PATH=/usr/local/include
           export PATH=$PATH:$GOPATH/bin
-
+          export CC=/opt/homebrew/opt/llvm/bin/clang
           make dist
       - uses: actions/upload-artifact@v4
         with:
diff --git a/.github/workflows/secscan.yaml b/.github/workflows/secscan.yaml
index d9743d9e..08d7dfc6 100644
--- a/.github/workflows/secscan.yaml
+++ b/.github/workflows/secscan.yaml
@@ -18,7 +18,7 @@ jobs:
         if: ${{ github.actor != 'dependabot[bot]' }}
       - name: Run Gosec Security Scanner
         if: ${{ github.actor != 'dependabot[bot]' }}
-        uses: securego/gosec@master
+        uses: securego/gosec@v2.21.2
         with:
           # we let the report trigger content trigger a failure using the GitHub Security features.
           args: '-no-fail -fmt sarif -out results.sarif ./...'
diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
index e6efe77f..2af3fd00 100644
--- a/.github/workflows/test.yml
+++ b/.github/workflows/test.yml
@@ -214,12 +214,13 @@ jobs:
         run: go version
       - name: Dependencies
         run: |
-          brew install protobuf grpc make protoc-gen-go protoc-gen-go-grpc
+          brew install protobuf grpc make protoc-gen-go protoc-gen-go-grpc libomp llvm
           pip install --user --no-cache-dir grpcio-tools==1.64.1
       - name: Test
         run: |
           export C_INCLUDE_PATH=/usr/local/include
           export CPLUS_INCLUDE_PATH=/usr/local/include
+          export CC=/opt/homebrew/opt/llvm/bin/clang
           # Used to run the newer GNUMake version from brew that supports --output-sync
           export PATH="/opt/homebrew/opt/make/libexec/gnubin:$PATH"
           BUILD_TYPE="GITHUB_CI_HAS_BROKEN_METAL" CMAKE_ARGS="-DGGML_F16C=OFF -DGGML_AVX512=OFF -DGGML_AVX2=OFF -DGGML_FMA=OFF" make --jobs 4 --output-sync=target test
diff --git a/.github/workflows/update_swagger.yaml b/.github/workflows/update_swagger.yaml
index b59e78c0..fb4762f8 100644
--- a/.github/workflows/update_swagger.yaml
+++ b/.github/workflows/update_swagger.yaml
@@ -25,7 +25,7 @@ jobs:
         run: |
           make protogen-go swagger
       - name: Create Pull Request
-        uses: peter-evans/create-pull-request@v6
+        uses: peter-evans/create-pull-request@v7
         with:
           token: ${{ secrets.UPDATE_BOT_TOKEN }}
           push-to-fork: ci-forks/LocalAI
diff --git a/Dockerfile b/Dockerfile
index 9d651760..f08cb9a0 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -13,7 +13,7 @@ ARG TARGETARCH
 ARG TARGETVARIANT
 
 ENV DEBIAN_FRONTEND=noninteractive
-ENV EXTERNAL_GRPC_BACKENDS="coqui:/build/backend/python/coqui/run.sh,huggingface-embeddings:/build/backend/python/sentencetransformers/run.sh,transformers:/build/backend/python/transformers/run.sh,sentencetransformers:/build/backend/python/sentencetransformers/run.sh,rerankers:/build/backend/python/rerankers/run.sh,autogptq:/build/backend/python/autogptq/run.sh,bark:/build/backend/python/bark/run.sh,diffusers:/build/backend/python/diffusers/run.sh,exllama:/build/backend/python/exllama/run.sh,openvoice:/build/backend/python/openvoice/run.sh,vall-e-x:/build/backend/python/vall-e-x/run.sh,vllm:/build/backend/python/vllm/run.sh,mamba:/build/backend/python/mamba/run.sh,exllama2:/build/backend/python/exllama2/run.sh,transformers-musicgen:/build/backend/python/transformers-musicgen/run.sh,parler-tts:/build/backend/python/parler-tts/run.sh"
+ENV EXTERNAL_GRPC_BACKENDS="coqui:/build/backend/python/coqui/run.sh,huggingface-embeddings:/build/backend/python/sentencetransformers/run.sh,transformers:/build/backend/python/transformers/run.sh,sentencetransformers:/build/backend/python/sentencetransformers/run.sh,rerankers:/build/backend/python/rerankers/run.sh,autogptq:/build/backend/python/autogptq/run.sh,bark:/build/backend/python/bark/run.sh,diffusers:/build/backend/python/diffusers/run.sh,openvoice:/build/backend/python/openvoice/run.sh,vall-e-x:/build/backend/python/vall-e-x/run.sh,vllm:/build/backend/python/vllm/run.sh,mamba:/build/backend/python/mamba/run.sh,exllama2:/build/backend/python/exllama2/run.sh,transformers-musicgen:/build/backend/python/transformers-musicgen/run.sh,parler-tts:/build/backend/python/parler-tts/run.sh"
 
 
 RUN apt-get update && \
@@ -263,14 +263,20 @@ EOT
 # In most cases, builder is the image you should be using - however, this can save build time if one just needs to copy backend-assets/grpc/stablediffusion and nothing else.
 FROM builder-base AS builder-sd
 
-COPY . .
-COPY .git .
+# stablediffusion does not tolerate a newer version of abseil, copy only over enough elements to build it
+COPY Makefile .
+COPY go.mod .
+COPY go.sum .
+COPY backend/backend.proto ./backend/backend.proto
+COPY backend/go/image/stablediffusion ./backend/go/image/stablediffusion
+COPY pkg/grpc ./pkg/grpc
+COPY pkg/stablediffusion ./pkg/stablediffusion
+RUN git init
+RUN make sources/go-stable-diffusion
+RUN touch prepare-sources
 
-RUN make prepare
-
-
-# stablediffusion does not tolerate a newer version of abseil, build it first
-RUN GRPC_BACKENDS=backend-assets/grpc/stablediffusion make build
+# Actually build the backend
+RUN GRPC_BACKENDS=backend-assets/grpc/stablediffusion make backend-assets/grpc/stablediffusion
 
 ###################################
 ###################################
@@ -285,8 +291,20 @@ COPY --from=grpc /opt/grpc /usr/local
 # Rebuild with defaults backends
 WORKDIR /build
 
+COPY . .
+COPY .git .
+
+RUN make prepare
+
 ## Build the binary
-RUN make build
+## If it's CUDA, we want to skip some of the llama-compat backends to save space
+## We only leave the most CPU-optimized variant and the fallback for the cublas build
+## (both will use CUDA for the actual computation)
+RUN if [ "${BUILD_TYPE}" = "cublas" ]; then \
+        SKIP_GRPC_BACKEND="backend-assets/grpc/llama-cpp-avx backend-assets/grpc/llama-cpp-avx2" make build; \
+    else \
+        make build; \
+    fi
 
 RUN if [ ! -d "/build/sources/go-piper/piper-phonemize/pi/lib/" ]; then \
         mkdir -p /build/sources/go-piper/piper-phonemize/pi/lib/ \
@@ -400,9 +418,6 @@ RUN if [[ ( "${EXTRA_BACKENDS}" =~ "coqui" || -z "${EXTRA_BACKENDS}" ) && "$IMAG
     ; fi && \
     if [[ ( "${EXTRA_BACKENDS}" =~ "transformers-musicgen" || -z "${EXTRA_BACKENDS}" ) && "$IMAGE_TYPE" == "extras" ]]; then \
         make -C backend/python/transformers-musicgen \
-    ; fi && \
-    if [[ ( "${EXTRA_BACKENDS}" =~ "exllama1" || -z "${EXTRA_BACKENDS}" ) && "$IMAGE_TYPE" == "extras" ]]; then \
-        make -C backend/python/exllama \
     ; fi
 
 RUN if [[ ( "${EXTRA_BACKENDS}" =~ "vall-e-x" || -z "${EXTRA_BACKENDS}" ) && "$IMAGE_TYPE" == "extras" ]]; then \
diff --git a/Makefile b/Makefile
index 988a4fb7..f9fa5476 100644
--- a/Makefile
+++ b/Makefile
@@ -8,7 +8,7 @@ DETECT_LIBS?=true
 # llama.cpp versions
 GOLLAMA_REPO?=https://github.com/go-skynet/go-llama.cpp
 GOLLAMA_VERSION?=2b57a8ae43e4699d3dc5d1496a1ccd42922993be
-CPPLLAMA_VERSION?=2f3c1466ff46a2413b0e363a5005c46538186ee6
+CPPLLAMA_VERSION?=23e0d70bacaaca1429d365a44aa9e7434f17823b
 
 # go-rwkv version
 RWKV_REPO?=https://github.com/donomii/go-rwkv.cpp
@@ -16,7 +16,7 @@ RWKV_VERSION?=661e7ae26d442f5cfebd2a0881b44e8c55949ec6
 
 # whisper.cpp version
 WHISPER_REPO?=https://github.com/ggerganov/whisper.cpp
-WHISPER_CPP_VERSION?=d65786ea540a5aef21f67cacfa6f134097727780
+WHISPER_CPP_VERSION?=049b3a0e53c8a8e4c4576c06a1a4fccf0063a73f
 
 # bert.cpp version
 BERT_REPO?=https://github.com/go-skynet/go-bert.cpp
@@ -534,10 +534,10 @@ protogen-go-clean:
 	$(RM) bin/*
 
 .PHONY: protogen-python
-protogen-python: autogptq-protogen bark-protogen coqui-protogen diffusers-protogen exllama-protogen exllama2-protogen mamba-protogen rerankers-protogen sentencetransformers-protogen transformers-protogen parler-tts-protogen transformers-musicgen-protogen vall-e-x-protogen vllm-protogen openvoice-protogen
+protogen-python: autogptq-protogen bark-protogen coqui-protogen diffusers-protogen exllama2-protogen mamba-protogen rerankers-protogen sentencetransformers-protogen transformers-protogen parler-tts-protogen transformers-musicgen-protogen vall-e-x-protogen vllm-protogen openvoice-protogen
 
 .PHONY: protogen-python-clean
-protogen-python-clean: autogptq-protogen-clean bark-protogen-clean coqui-protogen-clean diffusers-protogen-clean exllama-protogen-clean exllama2-protogen-clean mamba-protogen-clean sentencetransformers-protogen-clean rerankers-protogen-clean transformers-protogen-clean transformers-musicgen-protogen-clean parler-tts-protogen-clean vall-e-x-protogen-clean vllm-protogen-clean openvoice-protogen-clean
+protogen-python-clean: autogptq-protogen-clean bark-protogen-clean coqui-protogen-clean diffusers-protogen-clean  exllama2-protogen-clean mamba-protogen-clean sentencetransformers-protogen-clean rerankers-protogen-clean transformers-protogen-clean transformers-musicgen-protogen-clean parler-tts-protogen-clean vall-e-x-protogen-clean vllm-protogen-clean openvoice-protogen-clean
 
 .PHONY: autogptq-protogen
 autogptq-protogen:
@@ -571,14 +571,6 @@ diffusers-protogen:
 diffusers-protogen-clean:
 	$(MAKE) -C backend/python/diffusers protogen-clean
 
-.PHONY: exllama-protogen
-exllama-protogen:
-	$(MAKE) -C backend/python/exllama protogen
-
-.PHONY: exllama-protogen-clean
-exllama-protogen-clean:
-	$(MAKE) -C backend/python/exllama protogen-clean
-
 .PHONY: exllama2-protogen
 exllama2-protogen:
 	$(MAKE) -C backend/python/exllama2 protogen
@@ -675,7 +667,6 @@ prepare-extra-conda-environments: protogen-python
 	$(MAKE) -C backend/python/parler-tts
 	$(MAKE) -C backend/python/vall-e-x
 	$(MAKE) -C backend/python/openvoice
-	$(MAKE) -C backend/python/exllama
 	$(MAKE) -C backend/python/exllama2
 
 prepare-test-extra: protogen-python
@@ -846,7 +837,7 @@ endif
 
 backend-assets/grpc/whisper: sources/whisper.cpp sources/whisper.cpp/libwhisper.a backend-assets/grpc
 	CGO_LDFLAGS="$(CGO_LDFLAGS) $(CGO_LDFLAGS_WHISPER)" C_INCLUDE_PATH="$(CURDIR)/sources/whisper.cpp/include:$(CURDIR)/sources/whisper.cpp/ggml/include" LIBRARY_PATH=$(CURDIR)/sources/whisper.cpp \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/whisper ./backend/go/transcribe/
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/whisper ./backend/go/transcribe/whisper
 ifneq ($(UPX),)
 	$(UPX) backend-assets/grpc/whisper
 endif
diff --git a/README.md b/README.md
index ce3289f9..5b9a2c43 100644
--- a/README.md
+++ b/README.md
@@ -40,7 +40,7 @@
 
 > :bulb: Get help - [❓FAQ](https://localai.io/faq/) [💭Discussions](https://github.com/go-skynet/LocalAI/discussions) [:speech_balloon: Discord](https://discord.gg/uJAeKSAGDy) [:book: Documentation website](https://localai.io/)
 >
-> [💻 Quickstart](https://localai.io/basics/getting_started/) [📣 News](https://localai.io/basics/news/) [ 🛫 Examples ](https://github.com/go-skynet/LocalAI/tree/master/examples/) [ 🖼️ Models ](https://localai.io/models/) [ 🚀 Roadmap ](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap)
+> [💻 Quickstart](https://localai.io/basics/getting_started/) [🖼️ Models](https://models.localai.io/) [🚀 Roadmap](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap) [🥽 Demo](https://demo.localai.io) [🌍 Explorer](https://explorer.localai.io) [🛫 Examples](https://github.com/go-skynet/LocalAI/tree/master/examples/) 
 
 [![tests](https://github.com/go-skynet/LocalAI/actions/workflows/test.yml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/test.yml)[![Build and Release](https://github.com/go-skynet/LocalAI/actions/workflows/release.yaml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/release.yaml)[![build container images](https://github.com/go-skynet/LocalAI/actions/workflows/image.yml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/image.yml)[![Bump dependencies](https://github.com/go-skynet/LocalAI/actions/workflows/bump_deps.yaml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/bump_deps.yaml)[![Artifact Hub](https://img.shields.io/endpoint?url=https://artifacthub.io/badge/repository/localai)](https://artifacthub.io/packages/search?repo=localai)
 
@@ -72,6 +72,7 @@ docker run -ti --name local-ai -p 8080:8080 localai/localai:latest-aio-cpu
 
 [Roadmap](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap)
 
+- Aug 2024:  🆕 FLUX-1, [P2P Explorer](https://explorer.localai.io)
 - July 2024: 🔥🔥 🆕 P2P Dashboard, LocalAI Federated mode and AI Swarms: https://github.com/mudler/LocalAI/pull/2723
 - June 2024: 🆕 You can browse now the model gallery without LocalAI! Check out https://models.localai.io
 - June 2024: Support for models from OCI registries: https://github.com/mudler/LocalAI/pull/2628
diff --git a/aio/intel/image-gen.yaml b/aio/intel/image-gen.yaml
index eb724c92..45fe6b62 100644
--- a/aio/intel/image-gen.yaml
+++ b/aio/intel/image-gen.yaml
@@ -1,6 +1,6 @@
 name: stablediffusion
 parameters:
-  model: runwayml/stable-diffusion-v1-5
+  model: Lykon/dreamshaper-8
 backend: diffusers
 step: 25
 f16: true
diff --git a/backend/backend.proto b/backend/backend.proto
index 0d3d5f7f..4a8f31a9 100644
--- a/backend/backend.proto
+++ b/backend/backend.proto
@@ -16,6 +16,7 @@ service Backend {
   rpc GenerateImage(GenerateImageRequest) returns (Result) {}
   rpc AudioTranscription(TranscriptRequest) returns (TranscriptResult) {}
   rpc TTS(TTSRequest) returns (Result) {}
+  rpc SoundGeneration(SoundGenerationRequest) returns (Result) {}
   rpc TokenizeString(PredictOptions) returns (TokenizationResponse) {}
   rpc Status(HealthMessage) returns (StatusResponse) {}
 
@@ -270,6 +271,17 @@ message TTSRequest {
   optional string language = 5;
 }
 
+message SoundGenerationRequest {
+  string text = 1;
+  string model = 2;
+  string dst = 3;
+  optional float duration = 4;
+  optional float temperature = 5;
+  optional bool sample = 6;
+  optional string src = 7;
+  optional int32 src_divisor = 8;
+}
+
 message TokenizationResponse {
   int32 length = 1;
   repeated int32 tokens = 2;
diff --git a/backend/cpp/llama/grpc-server.cpp b/backend/cpp/llama/grpc-server.cpp
index 57ab46fe..56d59d21 100644
--- a/backend/cpp/llama/grpc-server.cpp
+++ b/backend/cpp/llama/grpc-server.cpp
@@ -13,15 +13,15 @@
 #include <getopt.h>
 #include "clip.h"
 #include "llava.h"
+#include "log.h"
 #include "stb_image.h"
 #include "common.h"
 #include "json.hpp"
 #include "llama.h"
-#include "grammar-parser.h"
 #include "backend.pb.h"
 #include "backend.grpc.pb.h"
 #include "utils.hpp"
-
+#include "sampling.h"
 // include std::regex
 #include <cstddef>
 #include <thread>
@@ -203,8 +203,8 @@ struct llama_client_slot
     std::string stopping_word;
 
     // sampling
-    struct llama_sampling_params sparams;
-    llama_sampling_context *ctx_sampling = nullptr;
+    struct gpt_sampler_params sparams;
+    gpt_sampler *ctx_sampling = nullptr;
 
     int32_t ga_i = 0;   // group-attention state
     int32_t ga_n = 1;   // group-attention factor
@@ -449,7 +449,7 @@ struct llama_server_context
             LOG_INFO("Multi Modal Mode Enabled", {});
             clp_ctx = clip_model_load(params.mmproj.c_str(), /*verbosity=*/ 1);
             if(clp_ctx == nullptr) {
-                LOG_ERROR("unable to load clip model", {{"model", params.mmproj}});
+                LOG_ERR("unable to load clip model: %s", params.mmproj.c_str());
                 return false;
             }
 
@@ -463,7 +463,7 @@ struct llama_server_context
         ctx = llama_init.context;
         if (model == nullptr)
         {
-            LOG_ERROR("unable to load model", {{"model", params.model}});
+            LOG_ERR("unable to load model: %s", params.model.c_str());
             return false;
         }
 
@@ -471,7 +471,7 @@ struct llama_server_context
             const int n_embd_clip = clip_n_mmproj_embd(clp_ctx);
             const int n_embd_llm  = llama_n_embd(model);
             if (n_embd_clip != n_embd_llm) {
-                LOG_TEE("%s: embedding dim of the multimodal projector (%d) is not equal to that of LLaMA (%d). Make sure that you use the correct mmproj file.\n", __func__, n_embd_clip, n_embd_llm);
+                LOG("%s: embedding dim of the multimodal projector (%d) is not equal to that of LLaMA (%d). Make sure that you use the correct mmproj file.\n", __func__, n_embd_clip, n_embd_llm);
                 llama_free(ctx);
                 llama_free_model(model);
                 return false;
@@ -490,7 +490,7 @@ struct llama_server_context
         std::vector<char> buf(1);
         int res = llama_chat_apply_template(model, nullptr, chat, 1, true, buf.data(), buf.size());
         if (res < 0) {
-            LOG_ERROR("The chat template comes with this model is not yet supported, falling back to chatml. This may cause the model to output suboptimal responses", {});
+            LOG_ERR("The chat template comes with this model is not yet supported, falling back to chatml. This may cause the model to output suboptimal responses", __func__);
             sparams.chat_template = "<|im_start|>"; // llama_chat_apply_template only checks if <|im_start|> exist in the template
         }
     }
@@ -619,7 +619,7 @@ struct llama_server_context
 
     bool launch_slot_with_data(llama_client_slot* &slot, json data) {
         slot_params default_params;
-        llama_sampling_params default_sparams;
+        gpt_sampler_params default_sparams;
  
         slot->params.stream             = json_value(data, "stream",            false);
         slot->params.cache_prompt       = json_value(data, "cache_prompt",      false);
@@ -628,7 +628,7 @@ struct llama_server_context
         slot->sparams.top_p             = json_value(data, "top_p",             default_sparams.top_p);
         slot->sparams.min_p             = json_value(data, "min_p",             default_sparams.min_p);
         slot->sparams.tfs_z             = json_value(data, "tfs_z",             default_sparams.tfs_z);
-        slot->sparams.typical_p         = json_value(data, "typical_p",         default_sparams.typical_p);
+        slot->sparams.typ_p             = json_value(data, "typical_p",         default_sparams.typ_p);
         slot->sparams.temp              = json_value(data, "temperature",       default_sparams.temp);
         slot->sparams.dynatemp_range    = json_value(data, "dynatemp_range",    default_sparams.dynatemp_range);
         slot->sparams.dynatemp_exponent = json_value(data, "dynatemp_exponent", default_sparams.dynatemp_exponent);
@@ -641,7 +641,7 @@ struct llama_server_context
         slot->sparams.mirostat_eta      = json_value(data, "mirostat_eta",      default_sparams.mirostat_eta);
         slot->sparams.penalize_nl       = json_value(data, "penalize_nl",       default_sparams.penalize_nl);
         slot->params.n_keep             = json_value(data, "n_keep",            slot->params.n_keep);
-        slot->params.seed               = json_value(data, "seed",              default_params.seed);
+        slot->sparams.seed               = json_value(data, "seed",              default_sparams.seed);
         slot->sparams.grammar           = json_value(data, "grammar",           default_sparams.grammar);
         slot->sparams.n_probs           = json_value(data, "n_probs",           default_sparams.n_probs);
         slot->sparams.min_keep          = json_value(data, "min_keep",          default_sparams.min_keep);
@@ -665,6 +665,7 @@ struct llama_server_context
             slot->params.input_prefix = "";
         }
 
+
         if (data.count("input_suffix") != 0)
         {
             slot->params.input_suffix = data["input_suffix"];
@@ -683,6 +684,10 @@ struct llama_server_context
             slot->prompt = "";
         }
 
+        if (json_value(data, "ignore_eos", false)) {
+                slot->sparams.logit_bias.push_back({llama_token_eos(model), -INFINITY});
+        }
+        /*
         slot->sparams.penalty_prompt_tokens.clear();
         slot->sparams.use_penalty_prompt_tokens = false;
         const auto &penalty_prompt = data.find("penalty_prompt");
@@ -718,14 +723,10 @@ struct llama_server_context
                 slot->sparams.use_penalty_prompt_tokens = true;
             }
         }
+      */
 
         slot->sparams.logit_bias.clear();
 
-        if (json_value(data, "ignore_eos", false))
-        {
-            slot->sparams.logit_bias[llama_token_eos(model)] = -INFINITY;
-        }
-
         const auto &logit_bias = data.find("logit_bias");
         if (logit_bias != data.end() && logit_bias->is_array())
         {
@@ -753,7 +754,7 @@ struct llama_server_context
                         llama_token tok = el[0].get<llama_token>();
                         if (tok >= 0 && tok < n_vocab)
                         {
-                            slot->sparams.logit_bias[tok] = bias;
+                            slot->sparams.logit_bias.push_back({tok, bias});
                         }
                     }
                     else if (el[0].is_string())
@@ -761,13 +762,13 @@ struct llama_server_context
                         auto toks = llama_tokenize(model, el[0].get<std::string>(), false);
                         for (auto tok : toks)
                         {
-                            slot->sparams.logit_bias[tok] = bias;
+                            slot->sparams.logit_bias.push_back({tok, bias});
                         }
                     }
                 }
             }
         }
-
+        
         slot->params.antiprompt.clear();
 
         const auto &stop = data.find("stop");
@@ -781,24 +782,22 @@ struct llama_server_context
                 }
             }
         }
-
-        const auto &samplers_sequence = data.find("samplers");
-        if (samplers_sequence != data.end() && samplers_sequence->is_array())
-        {
+        
+        const auto & samplers = data.find("samplers");
+        if (samplers != data.end() && samplers->is_array()) {
             std::vector<std::string> sampler_names;
-            for (const auto &sampler_name : *samplers_sequence)
-            {
-                if (sampler_name.is_string())
-                {
-                    sampler_names.emplace_back(sampler_name);
+                for (const auto & name : *samplers) {
+                    if (name.is_string()) {
+                        sampler_names.emplace_back(name);
+                    }
                 }
-            }
-            slot->sparams.samplers_sequence = llama_sampling_types_from_names(sampler_names, false);
+                slot->sparams.samplers = gpt_sampler_types_from_names(sampler_names, false);
         }
         else
         {
-            slot->sparams.samplers_sequence = default_sparams.samplers_sequence;
+                slot->sparams.samplers = default_sparams.samplers;
         }
+        
 
         if (multimodal)
         {
@@ -814,10 +813,11 @@ struct llama_server_context
                     img_sl.img_data = clip_image_u8_init();
                     if (!clip_image_load_from_bytes(image_buffer.data(), image_buffer.size(), img_sl.img_data))
                     {
-                        LOG_ERROR("failed to load image", {
-                            {"slot_id",   slot->id},
-                            {"img_sl_id", img_sl.id}
-                        });
+                        LOG_ERR("%s: failed to load image, slot_id: %d, img_sl_id: %d", 
+                             __func__,
+                             slot->id,
+                             img_sl.id
+                        );
                         return false;
                     }
                     LOG_VERBOSE("image loaded", {
@@ -855,12 +855,12 @@ struct llama_server_context
                                     }
                                 }
                                 if (!found) {
-                                    LOG_TEE("ERROR: Image with id: %i, not found.\n", img_id);
+                                    LOG("ERROR: Image with id: %i, not found.\n", img_id);
                                     slot->images.clear();
                                     return false;
                                 }
                             } catch (const std::invalid_argument& e) {
-                                LOG_TEE("Invalid image number id in prompt\n");
+                                LOG("Invalid image number id in prompt\n");
                                 slot->images.clear();
                                 return false;
                             }
@@ -875,10 +875,10 @@ struct llama_server_context
 
         if (slot->ctx_sampling != nullptr)
         {
-            llama_sampling_free(slot->ctx_sampling);
+            gpt_sampler_free(slot->ctx_sampling);
         }
-        slot->ctx_sampling = llama_sampling_init(slot->sparams);
-        llama_set_rng_seed(ctx, slot->params.seed);
+        slot->ctx_sampling = gpt_sampler_init(model, slot->sparams);
+        //llama_set_rng_seed(ctx, slot->params.seed);
         slot->command = LOAD_PROMPT;
 
         all_slots_are_idle = false;
@@ -888,7 +888,7 @@ struct llama_server_context
             {"task_id", slot->task_id},
         });
 
-        LOG_TEE("sampling: \n%s\n", llama_sampling_print(slot->sparams).c_str());
+      //  LOG("sampling: \n%s\n", llama_sampling_print(slot->sparams).c_str());
 
         return true;
     }
@@ -928,7 +928,7 @@ struct llama_server_context
                 };
                 if (llama_decode(ctx, batch_view) != 0)
                 {
-                    LOG_TEE("%s: llama_decode() failed\n", __func__);
+                    LOG("%s: llama_decode() failed\n", __func__);
                     return;
                 }
             }
@@ -940,7 +940,7 @@ struct llama_server_context
             }
         }
 
-        LOG_TEE("system prompt updated\n");
+        LOG("system prompt updated\n");
         system_need_update = false;
     }
 
@@ -1006,11 +1006,13 @@ struct llama_server_context
         slot.generated_text += token_str;
         slot.has_next_token = true;
 
+/*
         if (slot.ctx_sampling->params.use_penalty_prompt_tokens && result.tok != -1)
         {
             // we can change penalty_prompt_tokens because it is always created from scratch each request
             slot.ctx_sampling->params.penalty_prompt_tokens.push_back(result.tok);
         }
+        */
 
         // check if there is incomplete UTF-8 character at the end
         bool incomplete = false;
@@ -1119,8 +1121,8 @@ struct llama_server_context
                 continue;
             }
 
-            if (!llava_image_embed_make_with_clip_img(clp_ctx, params.n_threads, img.img_data, &img.image_embedding, &img.image_tokens)) {
-                LOG_TEE("Error processing the given image");
+            if (!llava_image_embed_make_with_clip_img(clp_ctx, params.cpuparams.n_threads, img.img_data, &img.image_embedding, &img.image_tokens)) {
+                LOG("Error processing the given image");
                 return false;
             }
 
@@ -1132,7 +1134,7 @@ struct llama_server_context
 
     void send_error(task_server& task, const std::string &error)
     {
-        LOG_TEE("task %i - error: %s\n", task.id, error.c_str());
+        LOG("task %i - error: %s\n", task.id, error.c_str());
         task_result res;
         res.id = task.id;
         res.multitask_id = task.multitask_id;
@@ -1144,13 +1146,11 @@ struct llama_server_context
 
     json get_formated_generation(llama_client_slot &slot)
     {
-        const auto eos_bias = slot.sparams.logit_bias.find(llama_token_eos(model));
-        const bool ignore_eos = eos_bias != slot.sparams.logit_bias.end() &&
-                                eos_bias->second < 0.0f && std::isinf(eos_bias->second);
-        std::vector<std::string> samplers_sequence;
-        for (const auto &sampler_type : slot.sparams.samplers_sequence)
+        std::vector<std::string> samplers;
+        samplers.reserve(slot.sparams.samplers.size());
+        for (const auto & sampler : slot.sparams.samplers)
         {
-            samplers_sequence.emplace_back(llama_sampling_type_to_str(sampler_type));
+            samplers.emplace_back(gpt_sampler_type_to_str(sampler));
         }
 
         return json {
@@ -1165,13 +1165,11 @@ struct llama_server_context
             {"top_p",             slot.sparams.top_p},
             {"min_p",             slot.sparams.min_p},
             {"tfs_z",             slot.sparams.tfs_z},
-            {"typical_p",         slot.sparams.typical_p},
+            {"typical_p",         slot.sparams.typ_p},
             {"repeat_last_n",     slot.sparams.penalty_last_n},
             {"repeat_penalty",    slot.sparams.penalty_repeat},
             {"presence_penalty",  slot.sparams.penalty_present},
             {"frequency_penalty", slot.sparams.penalty_freq},
-            {"penalty_prompt_tokens", slot.sparams.penalty_prompt_tokens},
-            {"use_penalty_prompt_tokens", slot.sparams.use_penalty_prompt_tokens},
             {"mirostat",          slot.sparams.mirostat},
             {"mirostat_tau",      slot.sparams.mirostat_tau},
             {"mirostat_eta",      slot.sparams.mirostat_eta},
@@ -1179,13 +1177,13 @@ struct llama_server_context
             {"stop",              slot.params.antiprompt},
             {"n_predict",         slot.params.n_predict},
             {"n_keep",            params.n_keep},
-            {"ignore_eos",        ignore_eos},
+            {"ignore_eos",        slot.sparams.ignore_eos},
             {"stream",            slot.params.stream},
-            {"logit_bias",        slot.sparams.logit_bias},
+      //      {"logit_bias",        slot.sparams.logit_bias},
             {"n_probs",           slot.sparams.n_probs},
             {"min_keep",          slot.sparams.min_keep},
             {"grammar",           slot.sparams.grammar},
-            {"samplers",          samplers_sequence}
+            {"samplers",          samplers}
         };
     }
 
@@ -1375,7 +1373,7 @@ struct llama_server_context
                 };
                 if (llama_decode(ctx, batch_view))
                 {
-                    LOG_TEE("%s : failed to eval\n", __func__);
+                    LOG("%s : failed to eval\n", __func__);
                     return false;
                 }
             }
@@ -1393,7 +1391,7 @@ struct llama_server_context
                 llama_batch batch_img = { n_eval, nullptr, (img.image_embedding + i * n_embd), nullptr, nullptr, nullptr, nullptr, slot.n_past, 1, 0, };
                 if (llama_decode(ctx, batch_img))
                 {
-                    LOG_TEE("%s : failed to eval image\n", __func__);
+                    LOG("%s : failed to eval image\n", __func__);
                     return false;
                 }
                 slot.n_past += n_eval;
@@ -1576,7 +1574,7 @@ struct llama_server_context
                     slot.n_past = 0;
                     slot.truncated = false;
                     slot.has_next_token = true;
-                    LOG_TEE("Context exhausted. Slot %d released (%d tokens in cache)\n", slot.id, (int) slot.cache_tokens.size());
+                    LOG("Context exhausted. Slot %d released (%d tokens in cache)\n", slot.id, (int) slot.cache_tokens.size());
 
                     continue;
                     // END LOCALAI changes
@@ -1714,7 +1712,7 @@ struct llama_server_context
 
                     if (!slot.params.cache_prompt)
                     {
-                        llama_sampling_reset(slot.ctx_sampling);
+                        gpt_sampler_reset(slot.ctx_sampling);
 
                         slot.n_past = 0;
                         slot.n_past_se = 0;
@@ -1726,7 +1724,7 @@ struct llama_server_context
                         // push the prompt into the sampling context (do not apply grammar)
                         for (auto &token : prompt_tokens)
                         {
-                            llama_sampling_accept(slot.ctx_sampling, ctx, token, false);
+                            gpt_sampler_accept(slot.ctx_sampling, token, false);
                         }
 
                         slot.n_past = common_part(slot.cache_tokens, prompt_tokens);
@@ -1824,10 +1822,11 @@ struct llama_server_context
 
                     if (has_images && !ingest_images(slot, n_batch))
                     {
-                        LOG_ERROR("failed processing images", {
-                            "slot_id", slot.id,
-                            "task_id", slot.task_id,
-                        });
+                        LOG_ERR("%s: failed processing images Slot id : %d, Task id: %d", 
+                            __func__,
+                            slot.id,
+                            slot.task_id
+                        );
                         // FIXME @phymbert: to be properly tested
                         //  early returning without changing the slot state will block the slot for ever
                         // no one at the moment is checking the return value
@@ -1867,10 +1866,10 @@ struct llama_server_context
                         const int bd = (slot.ga_w / slot.ga_n) * (slot.ga_n - 1);
                         const int dd = (slot.ga_w / slot.ga_n) - ib * bd - slot.ga_w;
 
-                        LOG_TEE("\n");
-                        LOG_TEE("shift: [%6d, %6d] + %6d -> [%6d, %6d]\n", slot.ga_i, slot.n_past_se, ib * bd, slot.ga_i + ib * bd, slot.n_past_se + ib * bd);
-                        LOG_TEE("div:   [%6d, %6d] / %6d -> [%6d, %6d]\n", slot.ga_i + ib * bd, slot.ga_i + ib * bd + slot.ga_w, slot.ga_n, (slot.ga_i + ib * bd) / slot.ga_n, (slot.ga_i + ib * bd + slot.ga_w) / slot.ga_n);
-                        LOG_TEE("shift: [%6d, %6d] + %6d -> [%6d, %6d]\n", slot.ga_i + ib * bd + slot.ga_w, slot.n_past_se + ib * bd, dd, slot.ga_i + ib * bd + slot.ga_w + dd, slot.n_past_se + ib * bd + dd);
+                        LOG("\n");
+                        LOG("shift: [%6d, %6d] + %6d -> [%6d, %6d]\n", slot.ga_i, slot.n_past_se, ib * bd, slot.ga_i + ib * bd, slot.n_past_se + ib * bd);
+                        LOG("div:   [%6d, %6d] / %6d -> [%6d, %6d]\n", slot.ga_i + ib * bd, slot.ga_i + ib * bd + slot.ga_w, slot.ga_n, (slot.ga_i + ib * bd) / slot.ga_n, (slot.ga_i + ib * bd + slot.ga_w) / slot.ga_n);
+                        LOG("shift: [%6d, %6d] + %6d -> [%6d, %6d]\n", slot.ga_i + ib * bd + slot.ga_w, slot.n_past_se + ib * bd, dd, slot.ga_i + ib * bd + slot.ga_w + dd, slot.n_past_se + ib * bd + dd);
 
                         llama_kv_cache_seq_add(ctx, slot.id, slot.ga_i, slot.n_past_se, ib * bd);
                         llama_kv_cache_seq_div(ctx, slot.id, slot.ga_i + ib * bd, slot.ga_i + ib * bd + slot.ga_w,slot.ga_n);
@@ -1880,7 +1879,7 @@ struct llama_server_context
 
                         slot.ga_i += slot.ga_w / slot.ga_n;
 
-                        LOG_TEE("\nn_past_old = %d, n_past = %d, ga_i = %d\n\n", slot.n_past_se + bd, slot.n_past_se, slot.ga_i);
+                        LOG("\nn_past_old = %d, n_past = %d, ga_i = %d\n\n", slot.n_past_se + bd, slot.n_past_se, slot.ga_i);
                     }
                     slot.n_past_se += n_tokens;
                 }
@@ -1905,11 +1904,11 @@ struct llama_server_context
                 if (n_batch == 1 || ret < 0)
                 {
                     // if you get here, it means the KV cache is full - try increasing it via the context size
-                    LOG_TEE("%s : failed to decode the batch, n_batch = %d, ret = %d\n", __func__, n_batch, ret);
+                    LOG("%s : failed to decode the batch, n_batch = %d, ret = %d\n", __func__, n_batch, ret);
                     return false;
                 }
 
-                LOG_TEE("%s : failed to find free space in the KV cache, retrying with smaller n_batch = %d\n", __func__, n_batch / 2);
+                LOG("%s : failed to find free space in the KV cache, retrying with smaller n_batch = %d\n", __func__, n_batch / 2);
 
                 // retry with half the batch size to try to find a free slot in the KV cache
                 n_batch /= 2;
@@ -1934,9 +1933,9 @@ struct llama_server_context
                 }
 
                 completion_token_output result;
-                const llama_token id = llama_sampling_sample(slot.ctx_sampling, ctx, NULL, slot.i_batch - i);
+                const llama_token id = gpt_sampler_sample(slot.ctx_sampling, ctx, slot.i_batch - i);
 
-                llama_sampling_accept(slot.ctx_sampling, ctx, id, true);
+                gpt_sampler_accept(slot.ctx_sampling, id, true);
 
                 slot.n_decoded += 1;
                 if (slot.n_decoded == 1)
@@ -1946,19 +1945,14 @@ struct llama_server_context
                     metrics.on_prompt_eval(slot);
                 }
 
-                llama_token_data_array cur_p = { slot.ctx_sampling->cur.data(), slot.ctx_sampling->cur.size(), false };
                 result.tok = id;
+                const auto * cur_p = gpt_sampler_get_candidates(slot.ctx_sampling);
 
-                const int32_t n_probs = slot.sparams.n_probs;
-                if (slot.sparams.temp <= 0 && n_probs > 0)
-                {
-                    // for llama_sample_token_greedy we need to sort candidates
-                    llama_sample_softmax(ctx, &cur_p);
-                }
-
-                for (size_t i = 0; i < std::min(cur_p.size, (size_t)n_probs); ++i)
-                {
-                    result.probs.push_back({cur_p.data[i].id, cur_p.data[i].p});
+                for (size_t i = 0; i < (size_t) slot.sparams.n_probs; ++i) {
+                    result.probs.push_back({
+                        cur_p->data[i].id,
+                        i >= cur_p->size ? 0.0f : cur_p->data[i].p,
+                    });
                 }
 
                 if (!process_token(result, slot))
@@ -2210,7 +2204,7 @@ static void params_parse(const backend::ModelOptions* request,
     params.model_alias =  request->modelfile();
     params.n_ctx = request->contextsize();
     //params.memory_f16 = request->f16memory();
-    params.n_threads = request->threads();
+    params.cpuparams.n_threads = request->threads();
     params.n_gpu_layers = request->ngpulayers();
     params.n_batch = request->nbatch();
     // Set params.n_parallel by environment variable (LLAMA_PARALLEL), defaults to 1
diff --git a/backend/cpp/llama/patches/01-llava.patch b/backend/cpp/llama/patches/01-llava.patch
new file mode 100644
index 00000000..fa122da2
--- /dev/null
+++ b/backend/cpp/llama/patches/01-llava.patch
@@ -0,0 +1,13 @@
+diff --git a/examples/llava/clip.cpp b/examples/llava/clip.cpp
+index 342042ff..224db9b5 100644
+--- a/examples/llava/clip.cpp
++++ b/examples/llava/clip.cpp
+@@ -2419,7 +2419,7 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima
+             struct ggml_tensor * patches = ggml_graph_get_tensor(gf, "patches");
+             int* patches_data = (int*)malloc(ggml_nbytes(patches));
+             for (int i = 0; i < num_patches; i++) {
+-                patches_data[i] = i + 1;
++                patches_data[i] = i;
+             }
+             ggml_backend_tensor_set(patches, patches_data, 0, ggml_nbytes(patches));
+             free(patches_data);
\ No newline at end of file
diff --git a/backend/cpp/llama/prepare.sh b/backend/cpp/llama/prepare.sh
index 6c00f27c..4c8393b9 100644
--- a/backend/cpp/llama/prepare.sh
+++ b/backend/cpp/llama/prepare.sh
@@ -1,5 +1,12 @@
 #!/bin/bash
 
+## Patches
+## Apply patches from the `patches` directory
+for patch in $(ls patches); do
+    echo "Applying patch $patch"
+    patch -d llama.cpp/ -p1 < patches/$patch
+done 
+
 cp -r CMakeLists.txt llama.cpp/examples/grpc-server/
 cp -r grpc-server.cpp llama.cpp/examples/grpc-server/
 cp -rfv json.hpp llama.cpp/examples/grpc-server/
diff --git a/backend/cpp/llama/utils.hpp b/backend/cpp/llama/utils.hpp
index c5dafbf0..198b6f26 100644
--- a/backend/cpp/llama/utils.hpp
+++ b/backend/cpp/llama/utils.hpp
@@ -480,31 +480,4 @@ static inline std::vector<uint8_t> base64_decode(const std::string & encoded_str
     }
 
     return ret;
-}
-
-//
-// random string / id
-//
-
-static std::string random_string()
-{
-    static const std::string str("0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz");
-
-    std::random_device rd;
-    std::mt19937 generator(rd());
-
-    std::string result(32, ' ');
-
-    for (int i = 0; i < 32; ++i) {
-        result[i] = str[generator() % str.size()];
-    }
-
-    return result;
-}
-
-static std::string gen_chatcmplid()
-{
-    std::stringstream chatcmplid;
-    chatcmplid << "chatcmpl-" << random_string();
-    return chatcmplid.str();
 }
\ No newline at end of file
diff --git a/backend/go/transcribe/transcript.go b/backend/go/transcribe/transcript.go
deleted file mode 100644
index 6831167f..00000000
--- a/backend/go/transcribe/transcript.go
+++ /dev/null
@@ -1,104 +0,0 @@
-package main
-
-import (
-	"fmt"
-	"os"
-	"os/exec"
-	"path/filepath"
-
-	"github.com/ggerganov/whisper.cpp/bindings/go/pkg/whisper"
-	"github.com/go-audio/wav"
-	"github.com/mudler/LocalAI/core/schema"
-)
-
-func ffmpegCommand(args []string) (string, error) {
-	cmd := exec.Command("ffmpeg", args...) // Constrain this to ffmpeg to permit security scanner to see that the command is safe.
-	cmd.Env = os.Environ()
-	out, err := cmd.CombinedOutput()
-	return string(out), err
-}
-
-// AudioToWav converts audio to wav for transcribe.
-// TODO: use https://github.com/mccoyst/ogg?
-func audioToWav(src, dst string) error {
-	commandArgs := []string{"-i", src, "-format", "s16le", "-ar", "16000", "-ac", "1", "-acodec", "pcm_s16le", dst}
-	out, err := ffmpegCommand(commandArgs)
-	if err != nil {
-		return fmt.Errorf("error: %w out: %s", err, out)
-	}
-	return nil
-}
-
-func Transcript(model whisper.Model, audiopath, language string, translate bool, threads uint) (schema.TranscriptionResult, error) {
-	res := schema.TranscriptionResult{}
-
-	dir, err := os.MkdirTemp("", "whisper")
-	if err != nil {
-		return res, err
-	}
-	defer os.RemoveAll(dir)
-
-	convertedPath := filepath.Join(dir, "converted.wav")
-
-	if err := audioToWav(audiopath, convertedPath); err != nil {
-		return res, err
-	}
-
-	// Open samples
-	fh, err := os.Open(convertedPath)
-	if err != nil {
-		return res, err
-	}
-	defer fh.Close()
-
-	// Read samples
-	d := wav.NewDecoder(fh)
-	buf, err := d.FullPCMBuffer()
-	if err != nil {
-		return res, err
-	}
-
-	data := buf.AsFloat32Buffer().Data
-
-	// Process samples
-	context, err := model.NewContext()
-	if err != nil {
-		return res, err
-
-	}
-
-	context.SetThreads(threads)
-
-	if language != "" {
-		context.SetLanguage(language)
-	} else {
-		context.SetLanguage("auto")
-	}
-
-	if translate {
-		context.SetTranslate(true)
-	}
-
-	if err := context.Process(data, nil, nil); err != nil {
-		return res, err
-	}
-
-	for {
-		s, err := context.NextSegment()
-		if err != nil {
-			break
-		}
-
-		var tokens []int
-		for _, t := range s.Tokens {
-			tokens = append(tokens, t.Id)
-		}
-
-		segment := schema.Segment{Id: s.Num, Text: s.Text, Start: s.Start, End: s.End, Tokens: tokens}
-		res.Segments = append(res.Segments, segment)
-
-		res.Text += s.Text
-	}
-
-	return res, nil
-}
diff --git a/backend/go/transcribe/whisper.go b/backend/go/transcribe/whisper.go
deleted file mode 100644
index 61ae98e9..00000000
--- a/backend/go/transcribe/whisper.go
+++ /dev/null
@@ -1,26 +0,0 @@
-package main
-
-// This is a wrapper to statisfy the GRPC service interface
-// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
-import (
-	"github.com/ggerganov/whisper.cpp/bindings/go/pkg/whisper"
-	"github.com/mudler/LocalAI/core/schema"
-	"github.com/mudler/LocalAI/pkg/grpc/base"
-	pb "github.com/mudler/LocalAI/pkg/grpc/proto"
-)
-
-type Whisper struct {
-	base.SingleThread
-	whisper whisper.Model
-}
-
-func (sd *Whisper) Load(opts *pb.ModelOptions) error {
-	// Note: the Model here is a path to a directory containing the model files
-	w, err := whisper.New(opts.ModelFile)
-	sd.whisper = w
-	return err
-}
-
-func (sd *Whisper) AudioTranscription(opts *pb.TranscriptRequest) (schema.TranscriptionResult, error) {
-	return Transcript(sd.whisper, opts.Dst, opts.Language, opts.Translate, uint(opts.Threads))
-}
diff --git a/backend/go/transcribe/main.go b/backend/go/transcribe/whisper/main.go
similarity index 100%
rename from backend/go/transcribe/main.go
rename to backend/go/transcribe/whisper/main.go
diff --git a/backend/go/transcribe/whisper/whisper.go b/backend/go/transcribe/whisper/whisper.go
new file mode 100644
index 00000000..63416bb3
--- /dev/null
+++ b/backend/go/transcribe/whisper/whisper.go
@@ -0,0 +1,105 @@
+package main
+
+// This is a wrapper to statisfy the GRPC service interface
+// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
+import (
+	"os"
+	"path/filepath"
+
+	"github.com/ggerganov/whisper.cpp/bindings/go/pkg/whisper"
+	"github.com/go-audio/wav"
+	"github.com/mudler/LocalAI/pkg/grpc/base"
+	pb "github.com/mudler/LocalAI/pkg/grpc/proto"
+	"github.com/mudler/LocalAI/pkg/utils"
+)
+
+type Whisper struct {
+	base.SingleThread
+	whisper whisper.Model
+}
+
+func (sd *Whisper) Load(opts *pb.ModelOptions) error {
+	// Note: the Model here is a path to a directory containing the model files
+	w, err := whisper.New(opts.ModelFile)
+	sd.whisper = w
+	return err
+}
+
+func (sd *Whisper) AudioTranscription(opts *pb.TranscriptRequest) (pb.TranscriptResult, error) {
+
+	dir, err := os.MkdirTemp("", "whisper")
+	if err != nil {
+		return pb.TranscriptResult{}, err
+	}
+	defer os.RemoveAll(dir)
+
+	convertedPath := filepath.Join(dir, "converted.wav")
+
+	if err := utils.AudioToWav(opts.Dst, convertedPath); err != nil {
+		return pb.TranscriptResult{}, err
+	}
+
+	// Open samples
+	fh, err := os.Open(convertedPath)
+	if err != nil {
+		return pb.TranscriptResult{}, err
+	}
+	defer fh.Close()
+
+	// Read samples
+	d := wav.NewDecoder(fh)
+	buf, err := d.FullPCMBuffer()
+	if err != nil {
+		return pb.TranscriptResult{}, err
+	}
+
+	data := buf.AsFloat32Buffer().Data
+
+	// Process samples
+	context, err := sd.whisper.NewContext()
+	if err != nil {
+		return pb.TranscriptResult{}, err
+
+	}
+
+	context.SetThreads(uint(opts.Threads))
+
+	if opts.Language != "" {
+		context.SetLanguage(opts.Language)
+	} else {
+		context.SetLanguage("auto")
+	}
+
+	if opts.Translate {
+		context.SetTranslate(true)
+	}
+
+	if err := context.Process(data, nil, nil); err != nil {
+		return pb.TranscriptResult{}, err
+	}
+
+	segments := []*pb.TranscriptSegment{}
+	text := ""
+	for {
+		s, err := context.NextSegment()
+		if err != nil {
+			break
+		}
+
+		var tokens []int32
+		for _, t := range s.Tokens {
+			tokens = append(tokens, int32(t.Id))
+		}
+
+		segment := &pb.TranscriptSegment{Id: int32(s.Num), Text: s.Text, Start: int64(s.Start), End: int64(s.End), Tokens: tokens}
+		segments = append(segments, segment)
+
+		text += s.Text
+	}
+
+	return pb.TranscriptResult{
+		Segments: segments,
+		Text:     text,
+	}, nil
+
+}
diff --git a/backend/python/autogptq/requirements-intel.txt b/backend/python/autogptq/requirements-intel.txt
index 755e19d8..d5e0173e 100644
--- a/backend/python/autogptq/requirements-intel.txt
+++ b/backend/python/autogptq/requirements-intel.txt
@@ -2,4 +2,4 @@
 intel-extension-for-pytorch
 torch
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
+setuptools==75.1.0 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
diff --git a/backend/python/autogptq/requirements.txt b/backend/python/autogptq/requirements.txt
index 53946f23..150fcc1b 100644
--- a/backend/python/autogptq/requirements.txt
+++ b/backend/python/autogptq/requirements.txt
@@ -1,6 +1,6 @@
 accelerate
 auto-gptq==0.7.1
-grpcio==1.65.4
+grpcio==1.66.1
 protobuf
 certifi
 transformers
\ No newline at end of file
diff --git a/backend/python/bark/requirements-intel.txt b/backend/python/bark/requirements-intel.txt
index 9feb6eef..c0e4dcaa 100644
--- a/backend/python/bark/requirements-intel.txt
+++ b/backend/python/bark/requirements-intel.txt
@@ -3,6 +3,6 @@ intel-extension-for-pytorch
 torch
 torchaudio
 optimum[openvino]
-setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==75.1.0 # https://github.com/mudler/LocalAI/issues/2406
 transformers
 accelerate
\ No newline at end of file
diff --git a/backend/python/bark/requirements.txt b/backend/python/bark/requirements.txt
index 08bfaec3..6404b98e 100644
--- a/backend/python/bark/requirements.txt
+++ b/backend/python/bark/requirements.txt
@@ -1,4 +1,4 @@
 bark==0.1.5
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 certifi
\ No newline at end of file
diff --git a/backend/python/common/template/requirements.txt b/backend/python/common/template/requirements.txt
index 35173155..21610c1c 100644
--- a/backend/python/common/template/requirements.txt
+++ b/backend/python/common/template/requirements.txt
@@ -1,2 +1,2 @@
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
\ No newline at end of file
diff --git a/backend/python/coqui/requirements-intel.txt b/backend/python/coqui/requirements-intel.txt
index 002a55c3..c0e4dcaa 100644
--- a/backend/python/coqui/requirements-intel.txt
+++ b/backend/python/coqui/requirements-intel.txt
@@ -3,6 +3,6 @@ intel-extension-for-pytorch
 torch
 torchaudio
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==75.1.0 # https://github.com/mudler/LocalAI/issues/2406
 transformers
 accelerate
\ No newline at end of file
diff --git a/backend/python/coqui/requirements.txt b/backend/python/coqui/requirements.txt
index 6125f739..d7708363 100644
--- a/backend/python/coqui/requirements.txt
+++ b/backend/python/coqui/requirements.txt
@@ -1,4 +1,4 @@
 TTS==0.22.0
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 certifi
\ No newline at end of file
diff --git a/backend/python/diffusers/backend.py b/backend/python/diffusers/backend.py
index 8f420848..e7ad1cdd 100755
--- a/backend/python/diffusers/backend.py
+++ b/backend/python/diffusers/backend.py
@@ -168,7 +168,7 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
             if request.CFGScale != 0:
                 self.cfg_scale = request.CFGScale
 
-            clipmodel = "runwayml/stable-diffusion-v1-5"
+            clipmodel = "Lykon/dreamshaper-8"
             if request.CLIPModel != "":
                 clipmodel = request.CLIPModel
             clipsubfolder = "text_encoder"
diff --git a/backend/python/diffusers/requirements-intel.txt b/backend/python/diffusers/requirements-intel.txt
index 1cc2e2a2..566278a8 100644
--- a/backend/python/diffusers/requirements-intel.txt
+++ b/backend/python/diffusers/requirements-intel.txt
@@ -3,7 +3,7 @@ intel-extension-for-pytorch
 torch
 torchvision
 optimum[openvino]
-setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==75.1.0 # https://github.com/mudler/LocalAI/issues/2406
 diffusers
 opencv-python
 transformers
diff --git a/backend/python/diffusers/requirements.txt b/backend/python/diffusers/requirements.txt
index b4195fc5..043c7aba 100644
--- a/backend/python/diffusers/requirements.txt
+++ b/backend/python/diffusers/requirements.txt
@@ -1,5 +1,5 @@
 setuptools
-grpcio==1.65.4
+grpcio==1.66.1
 pillow
 protobuf
 certifi
diff --git a/backend/python/diffusers/test.py b/backend/python/diffusers/test.py
index b5e381ba..0e92538e 100644
--- a/backend/python/diffusers/test.py
+++ b/backend/python/diffusers/test.py
@@ -53,7 +53,7 @@ class TestBackendServicer(unittest.TestCase):
             self.setUp()
             with grpc.insecure_channel("localhost:50051") as channel:
                 stub = backend_pb2_grpc.BackendStub(channel)
-                response = stub.LoadModel(backend_pb2.ModelOptions(Model="runwayml/stable-diffusion-v1-5"))
+                response = stub.LoadModel(backend_pb2.ModelOptions(Model="Lykon/dreamshaper-8"))
                 self.assertTrue(response.success)
                 self.assertEqual(response.message, "Model loaded successfully")
         except Exception as err:
@@ -71,7 +71,7 @@ class TestBackendServicer(unittest.TestCase):
             self.setUp()
             with grpc.insecure_channel("localhost:50051") as channel:
                 stub = backend_pb2_grpc.BackendStub(channel)
-                response = stub.LoadModel(backend_pb2.ModelOptions(Model="runwayml/stable-diffusion-v1-5"))
+                response = stub.LoadModel(backend_pb2.ModelOptions(Model="Lykon/dreamshaper-8"))
                 print(response.message)
                 self.assertTrue(response.success)
                 image_req = backend_pb2.GenerateImageRequest(positive_prompt="cat", width=16,height=16, dst="test.jpg")
@@ -81,4 +81,4 @@ class TestBackendServicer(unittest.TestCase):
             print(err)
             self.fail("Image gen service failed")
         finally:
-            self.tearDown()
\ No newline at end of file
+            self.tearDown()
diff --git a/backend/python/exllama/.gitignore b/backend/python/exllama/.gitignore
deleted file mode 100644
index 1d3a0654..00000000
--- a/backend/python/exllama/.gitignore
+++ /dev/null
@@ -1 +0,0 @@
-source
\ No newline at end of file
diff --git a/backend/python/exllama/Makefile b/backend/python/exllama/Makefile
deleted file mode 100644
index e6a67881..00000000
--- a/backend/python/exllama/Makefile
+++ /dev/null
@@ -1,25 +0,0 @@
-export CONDA_ENV_PATH = "exllama.yml"
-
-.PHONY: exllama
-exllama: protogen
-	bash install.sh ${CONDA_ENV_PATH}
-
-.PHONY: run
-run: protogen
-	@echo "Running exllama..."
-	bash run.sh
-	@echo "exllama run."
-
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto
-
-.PHONY: clean
-clean: protogen-clean
-	$(RM) -r venv source __pycache__
\ No newline at end of file
diff --git a/backend/python/exllama/README.md b/backend/python/exllama/README.md
deleted file mode 100644
index f9ed5e9f..00000000
--- a/backend/python/exllama/README.md
+++ /dev/null
@@ -1,5 +0,0 @@
-# Creating a separate environment for the exllama project
-
-```
-make exllama
-```
\ No newline at end of file
diff --git a/backend/python/exllama/backend.py b/backend/python/exllama/backend.py
deleted file mode 100755
index 58d1392c..00000000
--- a/backend/python/exllama/backend.py
+++ /dev/null
@@ -1,159 +0,0 @@
-#!/usr/bin/env python3
-import grpc
-from concurrent import futures
-import time
-import backend_pb2
-import backend_pb2_grpc
-import argparse
-import signal
-import sys
-import os, glob
-
-from pathlib import Path
-import torch
-import torch.nn.functional as F
-from torch import version as torch_version
-
-from source.tokenizer import ExLlamaTokenizer
-from source.generator import ExLlamaGenerator
-from source.model import ExLlama, ExLlamaCache, ExLlamaConfig
-
-_ONE_DAY_IN_SECONDS = 60 * 60 * 24
-
-# If MAX_WORKERS are specified in the environment use it, otherwise default to 1
-MAX_WORKERS = int(os.environ.get('PYTHON_GRPC_MAX_WORKERS', '1'))
-
-# Implement the BackendServicer class with the service methods
-class BackendServicer(backend_pb2_grpc.BackendServicer):
-    def generate(self,prompt, max_new_tokens):
-        self.generator.end_beam_search()
-
-        # Tokenizing the input
-        ids = self.generator.tokenizer.encode(prompt)
-
-        self.generator.gen_begin_reuse(ids)
-        initial_len = self.generator.sequence[0].shape[0]
-        has_leading_space = False
-        decoded_text = ''
-        for i in range(max_new_tokens):
-            token = self.generator.gen_single_token()
-            if i == 0 and self.generator.tokenizer.tokenizer.IdToPiece(int(token)).startswith('▁'):
-                has_leading_space = True
-
-            decoded_text = self.generator.tokenizer.decode(self.generator.sequence[0][initial_len:])
-            if has_leading_space:
-                decoded_text = ' ' + decoded_text
-
-            if token.item() == self.generator.tokenizer.eos_token_id:
-                break
-        return decoded_text
-    def Health(self, request, context):
-        return backend_pb2.Reply(message=bytes("OK", 'utf-8'))
-    def LoadModel(self, request, context):
-        try:
-            # https://github.com/turboderp/exllama/blob/master/example_cfg.py
-            model_directory = request.ModelFile
-
-            # Locate files we need within that directory
-            tokenizer_path = os.path.join(model_directory, "tokenizer.model")
-            model_config_path = os.path.join(model_directory, "config.json")
-            st_pattern = os.path.join(model_directory, "*.safetensors")
-            model_path = glob.glob(st_pattern)[0]
-
-            # Create config, model, tokenizer and generator
-
-            config = ExLlamaConfig(model_config_path)               # create config from config.json
-            config.model_path = model_path                          # supply path to model weights file
-            if (request.ContextSize):
-                config.max_seq_len = request.ContextSize            # override max sequence length
-                config.max_attention_size = request.ContextSize**2  # Should be set to context_size^2. 
-                # https://github.com/turboderp/exllama/issues/220#issuecomment-1720324163
-
-            # Set Rope scaling.
-            if (request.RopeFreqScale):
-                # Alpha value for Rope scaling. 
-                # Higher value increases context but adds perplexity.
-                # alpha_value and compress_pos_emb are mutually exclusive.
-                # https://github.com/turboderp/exllama/issues/115
-                config.alpha_value = request.RopeFreqScale
-                config.calculate_rotary_embedding_base()
-
-            model = ExLlama(config)                                 # create ExLlama instance and load the weights
-            tokenizer = ExLlamaTokenizer(tokenizer_path)            # create tokenizer from tokenizer model file
-
-            cache = ExLlamaCache(model, batch_size = 2)             # create cache for inference
-            generator = ExLlamaGenerator(model, tokenizer, cache)   # create generator
-
-            self.generator= generator
-            self.model = model
-            self.tokenizer = tokenizer
-            self.cache = cache
-        except Exception as err:
-            return backend_pb2.Result(success=False, message=f"Unexpected {err=}, {type(err)=}")
-        return backend_pb2.Result(message="Model loaded successfully", success=True)
-
-    def Predict(self, request, context):
-        penalty = 1.15
-        if request.Penalty != 0.0:
-            penalty = request.Penalty
-        self.generator.settings.token_repetition_penalty_max = penalty
-        self.generator.settings.temperature = request.Temperature
-        self.generator.settings.top_k = request.TopK
-        self.generator.settings.top_p = request.TopP
-
-        tokens = 512
-        if request.Tokens != 0:
-            tokens = request.Tokens
-
-        if self.cache.batch_size == 1:
-            del self.cache
-            self.cache = ExLlamaCache(self.model, batch_size=2)
-            self.generator = ExLlamaGenerator(self.model, self.tokenizer, self.cache)
-
-        t = self.generate(request.Prompt, tokens)
-
-        # Remove prompt from response if present
-        if request.Prompt in t:
-            t = t.replace(request.Prompt, "")
-
-        return backend_pb2.Result(message=bytes(t, encoding='utf-8'))
-
-    def PredictStream(self, request, context):
-        # Implement PredictStream RPC
-        #for reply in some_data_generator():
-        #    yield reply
-        # Not implemented yet
-        return self.Predict(request, context)
-
-
-def serve(address):
-    server = grpc.server(futures.ThreadPoolExecutor(max_workers=MAX_WORKERS))
-    backend_pb2_grpc.add_BackendServicer_to_server(BackendServicer(), server)
-    server.add_insecure_port(address)
-    server.start()
-    print("Server started. Listening on: " + address, file=sys.stderr)
-
-    # Define the signal handler function
-    def signal_handler(sig, frame):
-        print("Received termination signal. Shutting down...")
-        server.stop(0)
-        sys.exit(0)
-
-    # Set the signal handlers for SIGINT and SIGTERM
-    signal.signal(signal.SIGINT, signal_handler)
-    signal.signal(signal.SIGTERM, signal_handler)
-
-    try:
-        while True:
-            time.sleep(_ONE_DAY_IN_SECONDS)
-    except KeyboardInterrupt:
-        server.stop(0)
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser(description="Run the gRPC server.")
-    parser.add_argument(
-        "--addr", default="localhost:50051", help="The address to bind the server to."
-    )
-    args = parser.parse_args()
-
-    serve(args.addr)
\ No newline at end of file
diff --git a/backend/python/exllama/install.sh b/backend/python/exllama/install.sh
deleted file mode 100755
index d33c4356..00000000
--- a/backend/python/exllama/install.sh
+++ /dev/null
@@ -1,13 +0,0 @@
-#!/bin/bash
-set -e
-
-LIMIT_TARGETS="cublas"
-
-source $(dirname $0)/../common/libbackend.sh
-
-installRequirements
-
-git clone https://github.com/turboderp/exllama $MY_DIR/source
-uv pip install ${BUILD_ISOLATION_FLAG} --requirement ${MY_DIR}/source/requirements.txt
-
-cp -v ./*py $MY_DIR/source/
diff --git a/backend/python/exllama/requirements-cpu.txt b/backend/python/exllama/requirements-cpu.txt
deleted file mode 100644
index bbcdc8cd..00000000
--- a/backend/python/exllama/requirements-cpu.txt
+++ /dev/null
@@ -1,3 +0,0 @@
-transformers
-accelerate
-torch
\ No newline at end of file
diff --git a/backend/python/exllama/requirements-cublas11.txt b/backend/python/exllama/requirements-cublas11.txt
deleted file mode 100644
index 1dfb5b98..00000000
--- a/backend/python/exllama/requirements-cublas11.txt
+++ /dev/null
@@ -1,4 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch
-transformers
-accelerate
\ No newline at end of file
diff --git a/backend/python/exllama/requirements-cublas12.txt b/backend/python/exllama/requirements-cublas12.txt
deleted file mode 100644
index 1ec544cd..00000000
--- a/backend/python/exllama/requirements-cublas12.txt
+++ /dev/null
@@ -1,3 +0,0 @@
-torch
-transformers
-accelerate
\ No newline at end of file
diff --git a/backend/python/exllama/requirements.txt b/backend/python/exllama/requirements.txt
deleted file mode 100644
index 99b81098..00000000
--- a/backend/python/exllama/requirements.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-grpcio==1.65.5
-protobuf
-certifi
-setuptools
\ No newline at end of file
diff --git a/backend/python/exllama/run.sh b/backend/python/exllama/run.sh
deleted file mode 100755
index 63119689..00000000
--- a/backend/python/exllama/run.sh
+++ /dev/null
@@ -1,7 +0,0 @@
-#!/bin/bash
-LIMIT_TARGETS="cublas"
-BACKEND_FILE="${MY_DIR}/source/backend.py"
-
-source $(dirname $0)/../common/libbackend.sh
-
-startBackend $@
\ No newline at end of file
diff --git a/backend/python/exllama/test.sh b/backend/python/exllama/test.sh
deleted file mode 100755
index 6940b066..00000000
--- a/backend/python/exllama/test.sh
+++ /dev/null
@@ -1,6 +0,0 @@
-#!/bin/bash
-set -e
-
-source $(dirname $0)/../common/libbackend.sh
-
-runUnittests
diff --git a/backend/python/exllama2/requirements.txt b/backend/python/exllama2/requirements.txt
index ce15b0b6..6fb018a0 100644
--- a/backend/python/exllama2/requirements.txt
+++ b/backend/python/exllama2/requirements.txt
@@ -1,4 +1,4 @@
-grpcio==1.65.4
+grpcio==1.66.1
 protobuf
 certifi
 wheel
diff --git a/backend/python/mamba/requirements.txt b/backend/python/mamba/requirements.txt
index 920971ce..8e1b0195 100644
--- a/backend/python/mamba/requirements.txt
+++ b/backend/python/mamba/requirements.txt
@@ -1,3 +1,3 @@
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 certifi
\ No newline at end of file
diff --git a/backend/python/openvoice/requirements-intel.txt b/backend/python/openvoice/requirements-intel.txt
index 25921f8f..cea7de0b 100644
--- a/backend/python/openvoice/requirements-intel.txt
+++ b/backend/python/openvoice/requirements-intel.txt
@@ -2,7 +2,7 @@
 intel-extension-for-pytorch
 torch
 optimum[openvino]
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 librosa==0.9.1
 faster-whisper==1.0.3
@@ -15,7 +15,7 @@ unidecode==1.3.7
 whisper-timestamped==1.15.4
 openai
 python-dotenv
-pypinyin==0.50.0
+pypinyin==0.53.0
 cn2an==0.5.22
 jieba==0.42.1
 gradio==4.38.1
diff --git a/backend/python/openvoice/requirements.txt b/backend/python/openvoice/requirements.txt
index 13ce9c28..b38805be 100644
--- a/backend/python/openvoice/requirements.txt
+++ b/backend/python/openvoice/requirements.txt
@@ -1,4 +1,4 @@
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 librosa
 faster-whisper
diff --git a/backend/python/parler-tts/requirements-after.txt b/backend/python/parler-tts/requirements-after.txt
index 63599411..09811bf4 100644
--- a/backend/python/parler-tts/requirements-after.txt
+++ b/backend/python/parler-tts/requirements-after.txt
@@ -1 +1,3 @@
-git+https://github.com/huggingface/parler-tts.git@8e465f1b5fcd223478e07175cb40494d19ffbe17
\ No newline at end of file
+git+https://github.com/huggingface/parler-tts.git@8e465f1b5fcd223478e07175cb40494d19ffbe17
+llvmlite==0.43.0
+numba==0.60.0
diff --git a/backend/python/parler-tts/requirements-hipblas.txt b/backend/python/parler-tts/requirements-hipblas.txt
index af9e820e..b8758537 100644
--- a/backend/python/parler-tts/requirements-hipblas.txt
+++ b/backend/python/parler-tts/requirements-hipblas.txt
@@ -1,5 +1,5 @@
 --extra-index-url https://download.pytorch.org/whl/rocm6.0
-torch
-torchaudio
+torch==2.3.0+rocm6.0
+torchaudio==2.3.0+rocm6.0
 transformers
-accelerate
\ No newline at end of file
+accelerate
diff --git a/backend/python/parler-tts/requirements-intel.txt b/backend/python/parler-tts/requirements-intel.txt
index 002a55c3..c0e4dcaa 100644
--- a/backend/python/parler-tts/requirements-intel.txt
+++ b/backend/python/parler-tts/requirements-intel.txt
@@ -3,6 +3,6 @@ intel-extension-for-pytorch
 torch
 torchaudio
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==75.1.0 # https://github.com/mudler/LocalAI/issues/2406
 transformers
 accelerate
\ No newline at end of file
diff --git a/backend/python/parler-tts/requirements.txt b/backend/python/parler-tts/requirements.txt
index 1f17c892..0da3da13 100644
--- a/backend/python/parler-tts/requirements.txt
+++ b/backend/python/parler-tts/requirements.txt
@@ -1,4 +1,4 @@
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 certifi
 llvmlite==0.43.0
\ No newline at end of file
diff --git a/backend/python/rerankers/requirements-intel.txt b/backend/python/rerankers/requirements-intel.txt
index 1a39cf4f..e6bb4cc7 100644
--- a/backend/python/rerankers/requirements-intel.txt
+++ b/backend/python/rerankers/requirements-intel.txt
@@ -5,4 +5,4 @@ accelerate
 torch
 rerankers[transformers]
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
+setuptools==75.1.0 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
diff --git a/backend/python/rerankers/requirements.txt b/backend/python/rerankers/requirements.txt
index 2a8d18b1..8e1b0195 100644
--- a/backend/python/rerankers/requirements.txt
+++ b/backend/python/rerankers/requirements.txt
@@ -1,3 +1,3 @@
-grpcio==1.65.4
+grpcio==1.66.1
 protobuf
 certifi
\ No newline at end of file
diff --git a/backend/python/sentencetransformers/backend.py b/backend/python/sentencetransformers/backend.py
index 905015e1..2a20bf60 100755
--- a/backend/python/sentencetransformers/backend.py
+++ b/backend/python/sentencetransformers/backend.py
@@ -55,7 +55,7 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
         """
         model_name = request.Model
         try:
-            self.model = SentenceTransformer(model_name)
+            self.model = SentenceTransformer(model_name, trust_remote_code=request.TrustRemoteCode)
         except Exception as err:
             return backend_pb2.Result(success=False, message=f"Unexpected {err=}, {type(err)=}")
 
diff --git a/backend/python/sentencetransformers/requirements-cpu.txt b/backend/python/sentencetransformers/requirements-cpu.txt
index cd9924ef..f88de1e4 100644
--- a/backend/python/sentencetransformers/requirements-cpu.txt
+++ b/backend/python/sentencetransformers/requirements-cpu.txt
@@ -2,5 +2,5 @@ torch
 accelerate
 transformers
 bitsandbytes
-sentence-transformers==3.0.1
+sentence-transformers==3.1.0
 transformers
\ No newline at end of file
diff --git a/backend/python/sentencetransformers/requirements-cublas11.txt b/backend/python/sentencetransformers/requirements-cublas11.txt
index 1131f066..57caf1a1 100644
--- a/backend/python/sentencetransformers/requirements-cublas11.txt
+++ b/backend/python/sentencetransformers/requirements-cublas11.txt
@@ -1,5 +1,5 @@
 --extra-index-url https://download.pytorch.org/whl/cu118
 torch
 accelerate
-sentence-transformers==3.0.1
+sentence-transformers==3.1.0
 transformers
\ No newline at end of file
diff --git a/backend/python/sentencetransformers/requirements-cublas12.txt b/backend/python/sentencetransformers/requirements-cublas12.txt
index 2936e17b..834fa6a4 100644
--- a/backend/python/sentencetransformers/requirements-cublas12.txt
+++ b/backend/python/sentencetransformers/requirements-cublas12.txt
@@ -1,4 +1,4 @@
 torch
 accelerate
-sentence-transformers==3.0.1
+sentence-transformers==3.1.0
 transformers
\ No newline at end of file
diff --git a/backend/python/sentencetransformers/requirements-hipblas.txt b/backend/python/sentencetransformers/requirements-hipblas.txt
index 3b187c68..98a0a41b 100644
--- a/backend/python/sentencetransformers/requirements-hipblas.txt
+++ b/backend/python/sentencetransformers/requirements-hipblas.txt
@@ -1,5 +1,5 @@
 --extra-index-url https://download.pytorch.org/whl/rocm6.0
 torch
 accelerate
-sentence-transformers==3.0.1
+sentence-transformers==3.1.0
 transformers
\ No newline at end of file
diff --git a/backend/python/sentencetransformers/requirements-intel.txt b/backend/python/sentencetransformers/requirements-intel.txt
index 806e3d47..5948910d 100644
--- a/backend/python/sentencetransformers/requirements-intel.txt
+++ b/backend/python/sentencetransformers/requirements-intel.txt
@@ -4,5 +4,5 @@ torch
 optimum[openvino]
 setuptools==69.5.1 # https://github.com/mudler/LocalAI/issues/2406
 accelerate
-sentence-transformers==3.0.1
+sentence-transformers==3.1.0
 transformers
\ No newline at end of file
diff --git a/backend/python/sentencetransformers/requirements.txt b/backend/python/sentencetransformers/requirements.txt
index 920971ce..b9cb6061 100644
--- a/backend/python/sentencetransformers/requirements.txt
+++ b/backend/python/sentencetransformers/requirements.txt
@@ -1,3 +1,5 @@
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
-certifi
\ No newline at end of file
+certifi
+datasets
+einops
\ No newline at end of file
diff --git a/backend/python/transformers-musicgen/backend.py b/backend/python/transformers-musicgen/backend.py
index d41d9a5c..b9f1facf 100644
--- a/backend/python/transformers-musicgen/backend.py
+++ b/backend/python/transformers-musicgen/backend.py
@@ -15,7 +15,7 @@ import backend_pb2_grpc
 
 import grpc
 
-from scipy.io.wavfile import write as write_wav
+from scipy.io import wavfile
 from transformers import AutoProcessor, MusicgenForConditionalGeneration
 
 _ONE_DAY_IN_SECONDS = 60 * 60 * 24
@@ -63,6 +63,61 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
 
         return backend_pb2.Result(message="Model loaded successfully", success=True)
 
+    def SoundGeneration(self, request, context):
+        model_name = request.model
+        if model_name == "":
+            return backend_pb2.Result(success=False, message="request.model is required")
+        try:
+            self.processor = AutoProcessor.from_pretrained(model_name)
+            self.model = MusicgenForConditionalGeneration.from_pretrained(model_name)
+            inputs = None
+            if request.text == "":
+                inputs = self.model.get_unconditional_inputs(num_samples=1)
+            elif request.HasField('src'):
+                # TODO SECURITY CODE GOES HERE LOL
+                # WHO KNOWS IF THIS WORKS???
+                sample_rate, wsamples = wavfile.read('path_to_your_file.wav')
+                
+                if request.HasField('src_divisor'):
+                    wsamples = wsamples[: len(wsamples) // request.src_divisor]
+                
+                inputs = self.processor(
+                    audio=wsamples,
+                    sampling_rate=sample_rate,
+                    text=[request.text],
+                    padding=True,
+                    return_tensors="pt",
+                )
+            else:
+                inputs = self.processor(
+                    text=[request.text],
+                    padding=True,
+                    return_tensors="pt",
+                )
+            
+            tokens = 256
+            if request.HasField('duration'):
+                tokens = int(request.duration * 51.2) # 256 tokens = 5 seconds, therefore 51.2 tokens is one second
+            guidance = 3.0
+            if request.HasField('temperature'):
+                guidance = request.temperature
+            dosample = True
+            if request.HasField('sample'):
+                dosample = request.sample
+            audio_values = self.model.generate(**inputs, do_sample=dosample, guidance_scale=guidance, max_new_tokens=tokens)
+            print("[transformers-musicgen] SoundGeneration generated!", file=sys.stderr)
+            sampling_rate = self.model.config.audio_encoder.sampling_rate
+            wavfile.write(request.dst, rate=sampling_rate, data=audio_values[0, 0].numpy())
+            print("[transformers-musicgen] SoundGeneration saved to", request.dst, file=sys.stderr)
+            print("[transformers-musicgen] SoundGeneration for", file=sys.stderr)
+            print("[transformers-musicgen] SoundGeneration requested tokens", tokens, file=sys.stderr)
+            print(request, file=sys.stderr)
+        except Exception as err:
+            return backend_pb2.Result(success=False, message=f"Unexpected {err=}, {type(err)=}")
+        return backend_pb2.Result(success=True)
+
+
+# The TTS endpoint is older, and provides fewer features, but exists for compatibility reasons
     def TTS(self, request, context):
         model_name = request.model
         if model_name == "":
@@ -75,8 +130,7 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
                 padding=True,
                 return_tensors="pt",
             )
-            tokens = 256
-            # TODO get tokens from request?
+            tokens = 512 # No good place to set the "length" in TTS, so use 10s as a sane default
             audio_values = self.model.generate(**inputs, max_new_tokens=tokens)
             print("[transformers-musicgen] TTS generated!", file=sys.stderr)
             sampling_rate = self.model.config.audio_encoder.sampling_rate
diff --git a/backend/python/transformers-musicgen/requirements-intel.txt b/backend/python/transformers-musicgen/requirements-intel.txt
index 89bfa6a2..608d6939 100644
--- a/backend/python/transformers-musicgen/requirements-intel.txt
+++ b/backend/python/transformers-musicgen/requirements-intel.txt
@@ -4,4 +4,4 @@ transformers
 accelerate
 torch
 optimum[openvino]
-setuptools==69.5.1 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
+setuptools==75.1.0 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
diff --git a/backend/python/transformers-musicgen/requirements.txt b/backend/python/transformers-musicgen/requirements.txt
index a0076112..fb1119a9 100644
--- a/backend/python/transformers-musicgen/requirements.txt
+++ b/backend/python/transformers-musicgen/requirements.txt
@@ -1,4 +1,4 @@
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 scipy==1.14.0
 certifi
\ No newline at end of file
diff --git a/backend/python/transformers-musicgen/test.py b/backend/python/transformers-musicgen/test.py
index 777b399a..295de65e 100644
--- a/backend/python/transformers-musicgen/test.py
+++ b/backend/python/transformers-musicgen/test.py
@@ -63,7 +63,7 @@ class TestBackendServicer(unittest.TestCase):
 
     def test_tts(self):
         """
-        This method tests if the embeddings are generated successfully
+        This method tests if TTS is generated successfully
         """
         try:
             self.setUp()
@@ -77,5 +77,24 @@ class TestBackendServicer(unittest.TestCase):
         except Exception as err:
             print(err)
             self.fail("TTS service failed")
+        finally:
+            self.tearDown()
+
+    def test_sound_generation(self):
+        """
+        This method tests if SoundGeneration is generated successfully
+        """
+        try:
+            self.setUp()
+            with grpc.insecure_channel("localhost:50051") as channel:
+                stub = backend_pb2_grpc.BackendStub(channel)
+                response = stub.LoadModel(backend_pb2.ModelOptions(Model="facebook/musicgen-small"))
+                self.assertTrue(response.success)
+                sg_request = backend_pb2.SoundGenerationRequest(text="80s TV news production music hit for tonight's biggest story")
+                sg_response = stub.SoundGeneration(sg_request)
+                self.assertIsNotNone(sg_response)
+        except Exception as err:
+            print(err)
+            self.fail("SoundGeneration service failed")
         finally:
             self.tearDown()
\ No newline at end of file
diff --git a/backend/python/transformers/requirements.txt b/backend/python/transformers/requirements.txt
index 5531ea0e..b19c59c0 100644
--- a/backend/python/transformers/requirements.txt
+++ b/backend/python/transformers/requirements.txt
@@ -1,4 +1,4 @@
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 certifi
 setuptools==69.5.1 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
diff --git a/backend/python/vall-e-x/requirements-intel.txt b/backend/python/vall-e-x/requirements-intel.txt
index 6185314f..adbabeac 100644
--- a/backend/python/vall-e-x/requirements-intel.txt
+++ b/backend/python/vall-e-x/requirements-intel.txt
@@ -4,4 +4,4 @@ accelerate
 torch
 torchaudio
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
+setuptools==75.1.0 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
diff --git a/backend/python/vall-e-x/requirements.txt b/backend/python/vall-e-x/requirements.txt
index 920971ce..8e1b0195 100644
--- a/backend/python/vall-e-x/requirements.txt
+++ b/backend/python/vall-e-x/requirements.txt
@@ -1,3 +1,3 @@
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 certifi
\ No newline at end of file
diff --git a/backend/python/vllm/backend.py b/backend/python/vllm/backend.py
index 2d8b55db..2cf15c1c 100644
--- a/backend/python/vllm/backend.py
+++ b/backend/python/vllm/backend.py
@@ -135,6 +135,26 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
         res = await gen.__anext__()
         return res
 
+    def Embedding(self, request, context):
+        """
+        A gRPC method that calculates embeddings for a given sentence.
+
+        Args:
+            request: An EmbeddingRequest object that contains the request parameters.
+            context: A grpc.ServicerContext object that provides information about the RPC.
+
+        Returns:
+            An EmbeddingResult object that contains the calculated embeddings.
+        """
+        print("Calculated embeddings for: " + request.Embeddings, file=sys.stderr)
+        outputs = self.model.encode(request.Embeddings)
+        # Check if we have one result at least
+        if len(outputs) == 0:
+            context.set_code(grpc.StatusCode.INVALID_ARGUMENT)
+            context.set_details("No embeddings were calculated.")
+            return backend_pb2.EmbeddingResult()
+        return backend_pb2.EmbeddingResult(embeddings=outputs[0].outputs.embedding)
+
     async def PredictStream(self, request, context):
         """
         Generates text based on the given prompt and sampling parameters, and streams the results.
diff --git a/backend/python/vllm/requirements-intel.txt b/backend/python/vllm/requirements-intel.txt
index 7903282e..1f82c46e 100644
--- a/backend/python/vllm/requirements-intel.txt
+++ b/backend/python/vllm/requirements-intel.txt
@@ -4,4 +4,4 @@ accelerate
 torch
 transformers
 optimum[openvino]
-setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
+setuptools==75.1.0 # https://github.com/mudler/LocalAI/issues/2406
\ No newline at end of file
diff --git a/backend/python/vllm/requirements.txt b/backend/python/vllm/requirements.txt
index 99b81098..b9c192d5 100644
--- a/backend/python/vllm/requirements.txt
+++ b/backend/python/vllm/requirements.txt
@@ -1,4 +1,4 @@
-grpcio==1.65.5
+grpcio==1.66.1
 protobuf
 certifi
 setuptools
\ No newline at end of file
diff --git a/backend/python/vllm/test.py b/backend/python/vllm/test.py
index 83fb2651..9f325b10 100644
--- a/backend/python/vllm/test.py
+++ b/backend/python/vllm/test.py
@@ -72,5 +72,28 @@ class TestBackendServicer(unittest.TestCase):
         except Exception as err:
             print(err)
             self.fail("text service failed")
+        finally:
+            self.tearDown()
+
+    def test_embedding(self):
+        """
+        This method tests if the embeddings are generated successfully
+        """
+        try:
+            self.setUp()
+            with grpc.insecure_channel("localhost:50051") as channel:
+                stub = backend_pb2_grpc.BackendStub(channel)
+                response = stub.LoadModel(backend_pb2.ModelOptions(Model="intfloat/e5-mistral-7b-instruct"))
+                self.assertTrue(response.success)
+                embedding_request = backend_pb2.PredictOptions(Embeddings="This is a test sentence.")
+                embedding_response = stub.Embedding(embedding_request)
+                self.assertIsNotNone(embedding_response.embeddings)
+                # assert that is a list of floats
+                self.assertIsInstance(embedding_response.embeddings, list)
+                # assert that the list is not empty
+                self.assertTrue(len(embedding_response.embeddings) > 0)
+        except Exception as err:
+            print(err)
+            self.fail("Embedding service failed")
         finally:
             self.tearDown()
\ No newline at end of file
diff --git a/core/backend/backend_suite_test.go b/core/backend/backend_suite_test.go
new file mode 100644
index 00000000..541c91f6
--- /dev/null
+++ b/core/backend/backend_suite_test.go
@@ -0,0 +1,13 @@
+package backend_test
+
+import (
+	"testing"
+
+	. "github.com/onsi/ginkgo/v2"
+	. "github.com/onsi/gomega"
+)
+
+func TestBackend(t *testing.T) {
+	RegisterFailHandler(Fail)
+	RunSpecs(t, "Backend test suite")
+}
diff --git a/core/backend/llm.go b/core/backend/llm.go
index 9268fbbc..2b4564a8 100644
--- a/core/backend/llm.go
+++ b/core/backend/llm.go
@@ -9,6 +9,8 @@ import (
 	"sync"
 	"unicode/utf8"
 
+	"github.com/rs/zerolog/log"
+
 	"github.com/mudler/LocalAI/core/config"
 	"github.com/mudler/LocalAI/core/schema"
 
@@ -87,7 +89,7 @@ func ModelInference(ctx context.Context, s string, messages []schema.Message, im
 			case string:
 				protoMessages[i].Content = ct
 			default:
-				return nil, fmt.Errorf("Unsupported type for schema.Message.Content for inference: %T", ct)
+				return nil, fmt.Errorf("unsupported type for schema.Message.Content for inference: %T", ct)
 			}
 		}
 	}
@@ -181,13 +183,37 @@ func Finetune(config config.BackendConfig, input, prediction string) string {
 		mu.Lock()
 		reg, ok := cutstrings[c]
 		if !ok {
-			cutstrings[c] = regexp.MustCompile(c)
+			r, err := regexp.Compile(c)
+			if err != nil {
+				log.Fatal().Err(err).Msg("failed to compile regex")
+			}
+			cutstrings[c] = r
 			reg = cutstrings[c]
 		}
 		mu.Unlock()
 		prediction = reg.ReplaceAllString(prediction, "")
 	}
 
+	// extract results from the response which can be for instance inside XML tags
+	var predResult string
+	for _, r := range config.ExtractRegex {
+		mu.Lock()
+		reg, ok := cutstrings[r]
+		if !ok {
+			regex, err := regexp.Compile(r)
+			if err != nil {
+				log.Fatal().Err(err).Msg("failed to compile regex")
+			}
+			cutstrings[r] = regex
+			reg = regex
+		}
+		mu.Unlock()
+		predResult += reg.FindString(prediction)
+	}
+	if predResult != "" {
+		prediction = predResult
+	}
+
 	for _, c := range config.TrimSpace {
 		prediction = strings.TrimSpace(strings.TrimPrefix(prediction, c))
 	}
diff --git a/core/backend/llm_test.go b/core/backend/llm_test.go
new file mode 100644
index 00000000..f7630702
--- /dev/null
+++ b/core/backend/llm_test.go
@@ -0,0 +1,109 @@
+package backend_test
+
+import (
+	. "github.com/mudler/LocalAI/core/backend"
+	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/schema"
+
+	. "github.com/onsi/ginkgo/v2"
+	. "github.com/onsi/gomega"
+)
+
+var _ = Describe("LLM tests", func() {
+	Context("Finetune LLM output", func() {
+		var (
+			testConfig config.BackendConfig
+			input      string
+			prediction string
+			result     string
+		)
+
+		BeforeEach(func() {
+			testConfig = config.BackendConfig{
+				PredictionOptions: schema.PredictionOptions{
+					Echo: false,
+				},
+				LLMConfig: config.LLMConfig{
+					Cutstrings:   []string{`<.*?>`},                  // Example regex for removing XML tags
+					ExtractRegex: []string{`<result>(.*?)</result>`}, // Example regex to extract from tags
+					TrimSpace:    []string{" ", "\n"},
+					TrimSuffix:   []string{".", "!"},
+				},
+			}
+		})
+
+		Context("when echo is enabled", func() {
+			BeforeEach(func() {
+				testConfig.Echo = true
+				input = "Hello"
+				prediction = "World"
+			})
+
+			It("should prepend input to prediction", func() {
+				result = Finetune(testConfig, input, prediction)
+				Expect(result).To(Equal("HelloWorld"))
+			})
+		})
+
+		Context("when echo is disabled", func() {
+			BeforeEach(func() {
+				testConfig.Echo = false
+				input = "Hello"
+				prediction = "World"
+			})
+
+			It("should not modify the prediction with input", func() {
+				result = Finetune(testConfig, input, prediction)
+				Expect(result).To(Equal("World"))
+			})
+		})
+
+		Context("when cutstrings regex is applied", func() {
+			BeforeEach(func() {
+				input = ""
+				prediction = "<div>Hello</div> World"
+			})
+
+			It("should remove substrings matching cutstrings regex", func() {
+				result = Finetune(testConfig, input, prediction)
+				Expect(result).To(Equal("Hello World"))
+			})
+		})
+
+		Context("when extract regex is applied", func() {
+			BeforeEach(func() {
+				input = ""
+				prediction = "<response><result>42</result></response>"
+			})
+
+			It("should extract substrings matching the extract regex", func() {
+				result = Finetune(testConfig, input, prediction)
+				Expect(result).To(Equal("42"))
+			})
+		})
+
+		Context("when trimming spaces", func() {
+			BeforeEach(func() {
+				input = ""
+				prediction = "   Hello World   "
+			})
+
+			It("should trim spaces from the prediction", func() {
+				result = Finetune(testConfig, input, prediction)
+				Expect(result).To(Equal("Hello World"))
+			})
+		})
+
+		Context("when trimming suffixes", func() {
+			BeforeEach(func() {
+				input = ""
+				prediction = "Hello World."
+			})
+
+			It("should trim suffixes from the prediction", func() {
+				result = Finetune(testConfig, input, prediction)
+				Expect(result).To(Equal("Hello World"))
+			})
+		})
+	})
+})
diff --git a/core/backend/soundgeneration.go b/core/backend/soundgeneration.go
new file mode 100644
index 00000000..abd5221b
--- /dev/null
+++ b/core/backend/soundgeneration.go
@@ -0,0 +1,74 @@
+package backend
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"path/filepath"
+
+	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/pkg/grpc/proto"
+	"github.com/mudler/LocalAI/pkg/model"
+	"github.com/mudler/LocalAI/pkg/utils"
+)
+
+func SoundGeneration(
+	backend string,
+	modelFile string,
+	text string,
+	duration *float32,
+	temperature *float32,
+	doSample *bool,
+	sourceFile *string,
+	sourceDivisor *int32,
+	loader *model.ModelLoader,
+	appConfig *config.ApplicationConfig,
+	backendConfig config.BackendConfig,
+) (string, *proto.Result, error) {
+	if backend == "" {
+		return "", nil, fmt.Errorf("backend is a required parameter")
+	}
+
+	grpcOpts := gRPCModelOpts(backendConfig)
+	opts := modelOpts(config.BackendConfig{}, appConfig, []model.Option{
+		model.WithBackendString(backend),
+		model.WithModel(modelFile),
+		model.WithContext(appConfig.Context),
+		model.WithAssetDir(appConfig.AssetsDestination),
+		model.WithLoadGRPCLoadModelOpts(grpcOpts),
+	})
+
+	soundGenModel, err := loader.BackendLoader(opts...)
+	if err != nil {
+		return "", nil, err
+	}
+
+	if soundGenModel == nil {
+		return "", nil, fmt.Errorf("could not load sound generation model")
+	}
+
+	if err := os.MkdirAll(appConfig.AudioDir, 0750); err != nil {
+		return "", nil, fmt.Errorf("failed creating audio directory: %s", err)
+	}
+
+	fileName := utils.GenerateUniqueFileName(appConfig.AudioDir, "sound_generation", ".wav")
+	filePath := filepath.Join(appConfig.AudioDir, fileName)
+
+	res, err := soundGenModel.SoundGeneration(context.Background(), &proto.SoundGenerationRequest{
+		Text:        text,
+		Model:       modelFile,
+		Dst:         filePath,
+		Sample:      doSample,
+		Duration:    duration,
+		Temperature: temperature,
+		Src:         sourceFile,
+		SrcDivisor:  sourceDivisor,
+	})
+
+	// return RPC error if any
+	if !res.Success {
+		return "", nil, fmt.Errorf(res.Message)
+	}
+
+	return filePath, res, err
+}
diff --git a/core/backend/transcript.go b/core/backend/transcript.go
index 0980288f..6ebc7c10 100644
--- a/core/backend/transcript.go
+++ b/core/backend/transcript.go
@@ -3,12 +3,13 @@ package backend
 import (
 	"context"
 	"fmt"
+	"time"
 
 	"github.com/mudler/LocalAI/core/config"
 	"github.com/mudler/LocalAI/core/schema"
 
 	"github.com/mudler/LocalAI/pkg/grpc/proto"
-	model "github.com/mudler/LocalAI/pkg/model"
+	"github.com/mudler/LocalAI/pkg/model"
 )
 
 func ModelTranscription(audio, language string, translate bool, ml *model.ModelLoader, backendConfig config.BackendConfig, appConfig *config.ApplicationConfig) (*schema.TranscriptionResult, error) {
@@ -21,19 +22,40 @@ func ModelTranscription(audio, language string, translate bool, ml *model.ModelL
 		model.WithAssetDir(appConfig.AssetsDestination),
 	})
 
-	whisperModel, err := ml.BackendLoader(opts...)
+	transcriptionModel, err := ml.BackendLoader(opts...)
 	if err != nil {
 		return nil, err
 	}
 
-	if whisperModel == nil {
-		return nil, fmt.Errorf("could not load whisper model")
+	if transcriptionModel == nil {
+		return nil, fmt.Errorf("could not load transcription model")
 	}
 
-	return whisperModel.AudioTranscription(context.Background(), &proto.TranscriptRequest{
+	r, err := transcriptionModel.AudioTranscription(context.Background(), &proto.TranscriptRequest{
 		Dst:       audio,
 		Language:  language,
 		Translate: translate,
 		Threads:   uint32(*backendConfig.Threads),
 	})
+	if err != nil {
+		return nil, err
+	}
+	tr := &schema.TranscriptionResult{
+		Text: r.Text,
+	}
+	for _, s := range r.Segments {
+		var tks []int
+		for _, t := range s.Tokens {
+			tks = append(tks, int(t))
+		}
+		tr.Segments = append(tr.Segments,
+			schema.Segment{
+				Text:   s.Text,
+				Id:     int(s.Id),
+				Start:  time.Duration(s.Start),
+				End:    time.Duration(s.End),
+				Tokens: tks,
+			})
+	}
+	return tr, err
 }
diff --git a/core/backend/tts.go b/core/backend/tts.go
index ced73e13..258882ae 100644
--- a/core/backend/tts.go
+++ b/core/backend/tts.go
@@ -9,31 +9,15 @@ import (
 	"github.com/mudler/LocalAI/core/config"
 
 	"github.com/mudler/LocalAI/pkg/grpc/proto"
-	model "github.com/mudler/LocalAI/pkg/model"
+	"github.com/mudler/LocalAI/pkg/model"
 	"github.com/mudler/LocalAI/pkg/utils"
 )
 
-func generateUniqueFileName(dir, baseName, ext string) string {
-	counter := 1
-	fileName := baseName + ext
-
-	for {
-		filePath := filepath.Join(dir, fileName)
-		_, err := os.Stat(filePath)
-		if os.IsNotExist(err) {
-			return fileName
-		}
-
-		counter++
-		fileName = fmt.Sprintf("%s_%d%s", baseName, counter, ext)
-	}
-}
-
 func ModelTTS(
 	backend,
 	text,
 	modelFile,
-	voice ,
+	voice,
 	language string,
 	loader *model.ModelLoader,
 	appConfig *config.ApplicationConfig,
@@ -66,7 +50,7 @@ func ModelTTS(
 		return "", nil, fmt.Errorf("failed creating audio directory: %s", err)
 	}
 
-	fileName := generateUniqueFileName(appConfig.AudioDir, "tts", ".wav")
+	fileName := utils.GenerateUniqueFileName(appConfig.AudioDir, "tts", ".wav")
 	filePath := filepath.Join(appConfig.AudioDir, fileName)
 
 	// If the model file is not empty, we pass it joined with the model path
@@ -88,12 +72,15 @@ func ModelTTS(
 	}
 
 	res, err := ttsModel.TTS(context.Background(), &proto.TTSRequest{
-		Text:  text,
-		Model: modelPath,
-		Voice: voice,
-		Dst:   filePath,
+		Text:     text,
+		Model:    modelPath,
+		Voice:    voice,
+		Dst:      filePath,
 		Language: &language,
 	})
+	if err != nil {
+		return "", nil, err
+	}
 
 	// return RPC error if any
 	if !res.Success {
diff --git a/core/cli/api/p2p.go b/core/cli/api/p2p.go
new file mode 100644
index 00000000..a2ecfe3f
--- /dev/null
+++ b/core/cli/api/p2p.go
@@ -0,0 +1,80 @@
+package cli_api
+
+import (
+	"context"
+	"fmt"
+	"net"
+	"os"
+	"strings"
+
+	"github.com/mudler/LocalAI/core/p2p"
+	"github.com/mudler/edgevpn/pkg/node"
+
+	"github.com/rs/zerolog/log"
+)
+
+func StartP2PStack(ctx context.Context, address, token, networkID string, federated bool) error {
+	var n *node.Node
+	// Here we are avoiding creating multiple nodes:
+	// - if the federated mode is enabled, we create a federated node and expose a service
+	// - exposing a service creates a node with specific options, and we don't want to create another node
+
+	// If the federated mode is enabled, we expose a service to the local instance running
+	// at r.Address
+	if federated {
+		_, port, err := net.SplitHostPort(address)
+		if err != nil {
+			return err
+		}
+
+		// Here a new node is created and started
+		// and a service is exposed by the node
+		node, err := p2p.ExposeService(ctx, "localhost", port, token, p2p.NetworkID(networkID, p2p.FederatedID))
+		if err != nil {
+			return err
+		}
+
+		if err := p2p.ServiceDiscoverer(ctx, node, token, p2p.NetworkID(networkID, p2p.FederatedID), nil, false); err != nil {
+			return err
+		}
+
+		n = node
+	}
+
+	// If the p2p mode is enabled, we start the service discovery
+	if token != "" {
+		// If a node wasn't created previously, create it
+		if n == nil {
+			node, err := p2p.NewNode(token)
+			if err != nil {
+				return err
+			}
+			err = node.Start(ctx)
+			if err != nil {
+				return fmt.Errorf("starting new node: %w", err)
+			}
+			n = node
+		}
+
+		// Attach a ServiceDiscoverer to the p2p node
+		log.Info().Msg("Starting P2P server discovery...")
+		if err := p2p.ServiceDiscoverer(ctx, n, token, p2p.NetworkID(networkID, p2p.WorkerID), func(serviceID string, node p2p.NodeData) {
+			var tunnelAddresses []string
+			for _, v := range p2p.GetAvailableNodes(p2p.NetworkID(networkID, p2p.WorkerID)) {
+				if v.IsOnline() {
+					tunnelAddresses = append(tunnelAddresses, v.TunnelAddress)
+				} else {
+					log.Info().Msgf("Node %s is offline", v.ID)
+				}
+			}
+			tunnelEnvVar := strings.Join(tunnelAddresses, ",")
+
+			os.Setenv("LLAMACPP_GRPC_SERVERS", tunnelEnvVar)
+			log.Debug().Msgf("setting LLAMACPP_GRPC_SERVERS to %s", tunnelEnvVar)
+		}, true); err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
diff --git a/core/cli/cli.go b/core/cli/cli.go
index 2073778d..aed75d8a 100644
--- a/core/cli/cli.go
+++ b/core/cli/cli.go
@@ -8,12 +8,13 @@ import (
 var CLI struct {
 	cliContext.Context `embed:""`
 
-	Run        RunCMD        `cmd:"" help:"Run LocalAI, this the default command if no other command is specified. Run 'local-ai run --help' for more information" default:"withargs"`
-	Federated  FederatedCLI  `cmd:"" help:"Run LocalAI in federated mode"`
-	Models     ModelsCMD     `cmd:"" help:"Manage LocalAI models and definitions"`
-	TTS        TTSCMD        `cmd:"" help:"Convert text to speech"`
-	Transcript TranscriptCMD `cmd:"" help:"Convert audio to text"`
-	Worker     worker.Worker `cmd:"" help:"Run workers to distribute workload (llama.cpp-only)"`
-	Util       UtilCMD       `cmd:"" help:"Utility commands"`
-	Explorer   ExplorerCMD   `cmd:"" help:"Run p2p explorer"`
+	Run             RunCMD             `cmd:"" help:"Run LocalAI, this the default command if no other command is specified. Run 'local-ai run --help' for more information" default:"withargs"`
+	Federated       FederatedCLI       `cmd:"" help:"Run LocalAI in federated mode"`
+	Models          ModelsCMD          `cmd:"" help:"Manage LocalAI models and definitions"`
+	TTS             TTSCMD             `cmd:"" help:"Convert text to speech"`
+	SoundGeneration SoundGenerationCMD `cmd:"" help:"Generates audio files from text or audio"`
+	Transcript      TranscriptCMD      `cmd:"" help:"Convert audio to text"`
+	Worker          worker.Worker      `cmd:"" help:"Run workers to distribute workload (llama.cpp-only)"`
+	Util            UtilCMD            `cmd:"" help:"Utility commands"`
+	Explorer        ExplorerCMD        `cmd:"" help:"Run p2p explorer"`
 }
diff --git a/core/cli/run.go b/core/cli/run.go
index c469f05f..afb7204c 100644
--- a/core/cli/run.go
+++ b/core/cli/run.go
@@ -3,11 +3,10 @@ package cli
 import (
 	"context"
 	"fmt"
-	"net"
-	"os"
 	"strings"
 	"time"
 
+	cli_api "github.com/mudler/LocalAI/core/cli/api"
 	cliContext "github.com/mudler/LocalAI/core/cli/context"
 	"github.com/mudler/LocalAI/core/config"
 	"github.com/mudler/LocalAI/core/http"
@@ -42,29 +41,34 @@ type RunCMD struct {
 	Threads     int  `env:"LOCALAI_THREADS,THREADS" short:"t" help:"Number of threads used for parallel computation. Usage of the number of physical cores in the system is suggested" group:"performance"`
 	ContextSize int  `env:"LOCALAI_CONTEXT_SIZE,CONTEXT_SIZE" default:"512" help:"Default context size for models" group:"performance"`
 
-	Address                string   `env:"LOCALAI_ADDRESS,ADDRESS" default:":8080" help:"Bind address for the API server" group:"api"`
-	CORS                   bool     `env:"LOCALAI_CORS,CORS" help:"" group:"api"`
-	CORSAllowOrigins       string   `env:"LOCALAI_CORS_ALLOW_ORIGINS,CORS_ALLOW_ORIGINS" group:"api"`
-	LibraryPath            string   `env:"LOCALAI_LIBRARY_PATH,LIBRARY_PATH" help:"Path to the library directory (for e.g. external libraries used by backends)" default:"/usr/share/local-ai/libs" group:"backends"`
-	CSRF                   bool     `env:"LOCALAI_CSRF" help:"Enables fiber CSRF middleware" group:"api"`
-	UploadLimit            int      `env:"LOCALAI_UPLOAD_LIMIT,UPLOAD_LIMIT" default:"15" help:"Default upload-limit in MB" group:"api"`
-	APIKeys                []string `env:"LOCALAI_API_KEY,API_KEY" help:"List of API Keys to enable API authentication. When this is set, all the requests must be authenticated with one of these API keys" group:"api"`
-	DisableWebUI           bool     `env:"LOCALAI_DISABLE_WEBUI,DISABLE_WEBUI" default:"false" help:"Disable webui" group:"api"`
-	DisablePredownloadScan bool     `env:"LOCALAI_DISABLE_PREDOWNLOAD_SCAN" help:"If true, disables the best-effort security scanner before downloading any files." group:"hardening" default:"false"`
-	OpaqueErrors           bool     `env:"LOCALAI_OPAQUE_ERRORS" default:"false" help:"If true, all error responses are replaced with blank 500 errors. This is intended only for hardening against information leaks and is normally not recommended." group:"hardening"`
-	Peer2Peer              bool     `env:"LOCALAI_P2P,P2P" name:"p2p" default:"false" help:"Enable P2P mode" group:"p2p"`
-	Peer2PeerToken         string   `env:"LOCALAI_P2P_TOKEN,P2P_TOKEN,TOKEN" name:"p2ptoken" help:"Token for P2P mode (optional)" group:"p2p"`
-	Peer2PeerNetworkID     string   `env:"LOCALAI_P2P_NETWORK_ID,P2P_NETWORK_ID" help:"Network ID for P2P mode, can be set arbitrarly by the user for grouping a set of instances" group:"p2p"`
-	ParallelRequests       bool     `env:"LOCALAI_PARALLEL_REQUESTS,PARALLEL_REQUESTS" help:"Enable backends to handle multiple requests in parallel if they support it (e.g.: llama.cpp or vllm)" group:"backends"`
-	SingleActiveBackend    bool     `env:"LOCALAI_SINGLE_ACTIVE_BACKEND,SINGLE_ACTIVE_BACKEND" help:"Allow only one backend to be run at a time" group:"backends"`
-	PreloadBackendOnly     bool     `env:"LOCALAI_PRELOAD_BACKEND_ONLY,PRELOAD_BACKEND_ONLY" default:"false" help:"Do not launch the API services, only the preloaded models / backends are started (useful for multi-node setups)" group:"backends"`
-	ExternalGRPCBackends   []string `env:"LOCALAI_EXTERNAL_GRPC_BACKENDS,EXTERNAL_GRPC_BACKENDS" help:"A list of external grpc backends" group:"backends"`
-	EnableWatchdogIdle     bool     `env:"LOCALAI_WATCHDOG_IDLE,WATCHDOG_IDLE" default:"false" help:"Enable watchdog for stopping backends that are idle longer than the watchdog-idle-timeout" group:"backends"`
-	WatchdogIdleTimeout    string   `env:"LOCALAI_WATCHDOG_IDLE_TIMEOUT,WATCHDOG_IDLE_TIMEOUT" default:"15m" help:"Threshold beyond which an idle backend should be stopped" group:"backends"`
-	EnableWatchdogBusy     bool     `env:"LOCALAI_WATCHDOG_BUSY,WATCHDOG_BUSY" default:"false" help:"Enable watchdog for stopping backends that are busy longer than the watchdog-busy-timeout" group:"backends"`
-	WatchdogBusyTimeout    string   `env:"LOCALAI_WATCHDOG_BUSY_TIMEOUT,WATCHDOG_BUSY_TIMEOUT" default:"5m" help:"Threshold beyond which a busy backend should be stopped" group:"backends"`
-	Federated              bool     `env:"LOCALAI_FEDERATED,FEDERATED" help:"Enable federated instance" group:"federated"`
-	DisableGalleryEndpoint bool     `env:"LOCALAI_DISABLE_GALLERY_ENDPOINT,DISABLE_GALLERY_ENDPOINT" help:"Disable the gallery endpoints" group:"api"`
+	Address                            string   `env:"LOCALAI_ADDRESS,ADDRESS" default:":8080" help:"Bind address for the API server" group:"api"`
+	CORS                               bool     `env:"LOCALAI_CORS,CORS" help:"" group:"api"`
+	CORSAllowOrigins                   string   `env:"LOCALAI_CORS_ALLOW_ORIGINS,CORS_ALLOW_ORIGINS" group:"api"`
+	LibraryPath                        string   `env:"LOCALAI_LIBRARY_PATH,LIBRARY_PATH" help:"Path to the library directory (for e.g. external libraries used by backends)" default:"/usr/share/local-ai/libs" group:"backends"`
+	CSRF                               bool     `env:"LOCALAI_CSRF" help:"Enables fiber CSRF middleware" group:"api"`
+	UploadLimit                        int      `env:"LOCALAI_UPLOAD_LIMIT,UPLOAD_LIMIT" default:"15" help:"Default upload-limit in MB" group:"api"`
+	APIKeys                            []string `env:"LOCALAI_API_KEY,API_KEY" help:"List of API Keys to enable API authentication. When this is set, all the requests must be authenticated with one of these API keys" group:"api"`
+	DisableWebUI                       bool     `env:"LOCALAI_DISABLE_WEBUI,DISABLE_WEBUI" default:"false" help:"Disable webui" group:"api"`
+	DisablePredownloadScan             bool     `env:"LOCALAI_DISABLE_PREDOWNLOAD_SCAN" help:"If true, disables the best-effort security scanner before downloading any files." group:"hardening" default:"false"`
+	OpaqueErrors                       bool     `env:"LOCALAI_OPAQUE_ERRORS" default:"false" help:"If true, all error responses are replaced with blank 500 errors. This is intended only for hardening against information leaks and is normally not recommended." group:"hardening"`
+	UseSubtleKeyComparison             bool     `env:"LOCALAI_SUBTLE_KEY_COMPARISON" default:"false" help:"If true, API Key validation comparisons will be performed using constant-time comparisons rather than simple equality. This trades off performance on each request for resiliancy against timing attacks." group:"hardening"`
+	DisableApiKeyRequirementForHttpGet bool     `env:"LOCALAI_DISABLE_API_KEY_REQUIREMENT_FOR_HTTP_GET" default:"false" help:"If true, a valid API key is not required to issue GET requests to portions of the web ui. This should only be enabled in secure testing environments" group:"hardening"`
+	HttpGetExemptedEndpoints           []string `env:"LOCALAI_HTTP_GET_EXEMPTED_ENDPOINTS" default:"^/$,^/browse/?$,^/talk/?$,^/p2p/?$,^/chat/?$,^/text2image/?$,^/tts/?$,^/static/.*$,^/swagger.*$" help:"If LOCALAI_DISABLE_API_KEY_REQUIREMENT_FOR_HTTP_GET is overriden to true, this is the list of endpoints to exempt. Only adjust this in case of a security incident or as a result of a personal security posture review" group:"hardening"`
+	Peer2Peer                          bool     `env:"LOCALAI_P2P,P2P" name:"p2p" default:"false" help:"Enable P2P mode" group:"p2p"`
+	Peer2PeerDHTInterval               int      `env:"LOCALAI_P2P_DHT_INTERVAL,P2P_DHT_INTERVAL" default:"360" name:"p2p-dht-interval" help:"Interval for DHT refresh (used during token generation)" group:"p2p"`
+	Peer2PeerOTPInterval               int      `env:"LOCALAI_P2P_OTP_INTERVAL,P2P_OTP_INTERVAL" default:"9000" name:"p2p-otp-interval" help:"Interval for OTP refresh (used during token generation)" group:"p2p"`
+	Peer2PeerToken                     string   `env:"LOCALAI_P2P_TOKEN,P2P_TOKEN,TOKEN" name:"p2ptoken" help:"Token for P2P mode (optional)" group:"p2p"`
+	Peer2PeerNetworkID                 string   `env:"LOCALAI_P2P_NETWORK_ID,P2P_NETWORK_ID" help:"Network ID for P2P mode, can be set arbitrarly by the user for grouping a set of instances" group:"p2p"`
+	ParallelRequests                   bool     `env:"LOCALAI_PARALLEL_REQUESTS,PARALLEL_REQUESTS" help:"Enable backends to handle multiple requests in parallel if they support it (e.g.: llama.cpp or vllm)" group:"backends"`
+	SingleActiveBackend                bool     `env:"LOCALAI_SINGLE_ACTIVE_BACKEND,SINGLE_ACTIVE_BACKEND" help:"Allow only one backend to be run at a time" group:"backends"`
+	PreloadBackendOnly                 bool     `env:"LOCALAI_PRELOAD_BACKEND_ONLY,PRELOAD_BACKEND_ONLY" default:"false" help:"Do not launch the API services, only the preloaded models / backends are started (useful for multi-node setups)" group:"backends"`
+	ExternalGRPCBackends               []string `env:"LOCALAI_EXTERNAL_GRPC_BACKENDS,EXTERNAL_GRPC_BACKENDS" help:"A list of external grpc backends" group:"backends"`
+	EnableWatchdogIdle                 bool     `env:"LOCALAI_WATCHDOG_IDLE,WATCHDOG_IDLE" default:"false" help:"Enable watchdog for stopping backends that are idle longer than the watchdog-idle-timeout" group:"backends"`
+	WatchdogIdleTimeout                string   `env:"LOCALAI_WATCHDOG_IDLE_TIMEOUT,WATCHDOG_IDLE_TIMEOUT" default:"15m" help:"Threshold beyond which an idle backend should be stopped" group:"backends"`
+	EnableWatchdogBusy                 bool     `env:"LOCALAI_WATCHDOG_BUSY,WATCHDOG_BUSY" default:"false" help:"Enable watchdog for stopping backends that are busy longer than the watchdog-busy-timeout" group:"backends"`
+	WatchdogBusyTimeout                string   `env:"LOCALAI_WATCHDOG_BUSY_TIMEOUT,WATCHDOG_BUSY_TIMEOUT" default:"5m" help:"Threshold beyond which a busy backend should be stopped" group:"backends"`
+	Federated                          bool     `env:"LOCALAI_FEDERATED,FEDERATED" help:"Enable federated instance" group:"federated"`
+	DisableGalleryEndpoint             bool     `env:"LOCALAI_DISABLE_GALLERY_ENDPOINT,DISABLE_GALLERY_ENDPOINT" help:"Disable the gallery endpoints" group:"api"`
 }
 
 func (r *RunCMD) Run(ctx *cliContext.Context) error {
@@ -96,6 +100,9 @@ func (r *RunCMD) Run(ctx *cliContext.Context) error {
 		config.WithModelsURL(append(r.Models, r.ModelArgs...)...),
 		config.WithOpaqueErrors(r.OpaqueErrors),
 		config.WithEnforcedPredownloadScans(!r.DisablePredownloadScan),
+		config.WithSubtleKeyComparison(r.UseSubtleKeyComparison),
+		config.WithDisableApiKeyRequirementForHttpGet(r.DisableApiKeyRequirementForHttpGet),
+		config.WithHttpGetExemptedEndpoints(r.HttpGetExemptedEndpoints),
 		config.WithP2PNetworkID(r.Peer2PeerNetworkID),
 	}
 
@@ -107,7 +114,7 @@ func (r *RunCMD) Run(ctx *cliContext.Context) error {
 			// IF no token is provided, and p2p is enabled,
 			// we generate one and wait for the user to pick up the token (this is for interactive)
 			log.Info().Msg("No token provided, generating one")
-			token = p2p.GenerateToken()
+			token = p2p.GenerateToken(r.Peer2PeerDHTInterval, r.Peer2PeerOTPInterval)
 			log.Info().Msg("Generated Token:")
 			fmt.Println(token)
 
@@ -115,45 +122,12 @@ func (r *RunCMD) Run(ctx *cliContext.Context) error {
 			fmt.Printf("export TOKEN=\"%s\"\nlocal-ai worker p2p-llama-cpp-rpc\n", token)
 		}
 		opts = append(opts, config.WithP2PToken(token))
-
-		node, err := p2p.NewNode(token)
-		if err != nil {
-			return err
-		}
-
-		log.Info().Msg("Starting P2P server discovery...")
-		if err := p2p.ServiceDiscoverer(context.Background(), node, token, p2p.NetworkID(r.Peer2PeerNetworkID, p2p.WorkerID), func(serviceID string, node p2p.NodeData) {
-			var tunnelAddresses []string
-			for _, v := range p2p.GetAvailableNodes(p2p.NetworkID(r.Peer2PeerNetworkID, p2p.WorkerID)) {
-				if v.IsOnline() {
-					tunnelAddresses = append(tunnelAddresses, v.TunnelAddress)
-				} else {
-					log.Info().Msgf("Node %s is offline", v.ID)
-				}
-			}
-			tunnelEnvVar := strings.Join(tunnelAddresses, ",")
-
-			os.Setenv("LLAMACPP_GRPC_SERVERS", tunnelEnvVar)
-			log.Debug().Msgf("setting LLAMACPP_GRPC_SERVERS to %s", tunnelEnvVar)
-		}, true); err != nil {
-			return err
-		}
 	}
 
-	if r.Federated {
-		_, port, err := net.SplitHostPort(r.Address)
-		if err != nil {
-			return err
-		}
-		fedCtx := context.Background()
-		node, err := p2p.ExposeService(fedCtx, "localhost", port, token, p2p.NetworkID(r.Peer2PeerNetworkID, p2p.FederatedID))
-		if err != nil {
-			return err
-		}
+	backgroundCtx := context.Background()
 
-		if err := p2p.ServiceDiscoverer(fedCtx, node, token, p2p.NetworkID(r.Peer2PeerNetworkID, p2p.FederatedID), nil, false); err != nil {
-			return err
-		}
+	if err := cli_api.StartP2PStack(backgroundCtx, r.Address, token, r.Peer2PeerNetworkID, r.Federated); err != nil {
+		return err
 	}
 
 	idleWatchDog := r.EnableWatchdogIdle
diff --git a/core/cli/soundgeneration.go b/core/cli/soundgeneration.go
new file mode 100644
index 00000000..5711b199
--- /dev/null
+++ b/core/cli/soundgeneration.go
@@ -0,0 +1,110 @@
+package cli
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"path/filepath"
+	"strconv"
+	"strings"
+
+	"github.com/mudler/LocalAI/core/backend"
+	cliContext "github.com/mudler/LocalAI/core/cli/context"
+	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/pkg/model"
+	"github.com/rs/zerolog/log"
+)
+
+type SoundGenerationCMD struct {
+	Text []string `arg:""`
+
+	Backend                string   `short:"b" required:"" help:"Backend to run the SoundGeneration model"`
+	Model                  string   `short:"m" required:"" help:"Model name to run the SoundGeneration"`
+	Duration               string   `short:"d" help:"If specified, the length of audio to generate in seconds"`
+	Temperature            string   `short:"t" help:"If specified, the temperature of the generation"`
+	InputFile              string   `short:"i" help:"If specified, the input file to condition generation upon"`
+	InputFileSampleDivisor string   `short:"f" help:"If InputFile and this divisor is specified, the first portion of the sample file will be used"`
+	DoSample               bool     `short:"s" default:"true" help:"Enables sampling from the model. Better quality at the cost of speed. Defaults to enabled."`
+	OutputFile             string   `short:"o" type:"path" help:"The path to write the output wav file"`
+	ModelsPath             string   `env:"LOCALAI_MODELS_PATH,MODELS_PATH" type:"path" default:"${basepath}/models" help:"Path containing models used for inferencing" group:"storage"`
+	BackendAssetsPath      string   `env:"LOCALAI_BACKEND_ASSETS_PATH,BACKEND_ASSETS_PATH" type:"path" default:"/tmp/localai/backend_data" help:"Path used to extract libraries that are required by some of the backends in runtime" group:"storage"`
+	ExternalGRPCBackends   []string `env:"LOCALAI_EXTERNAL_GRPC_BACKENDS,EXTERNAL_GRPC_BACKENDS" help:"A list of external grpc backends" group:"backends"`
+}
+
+func parseToFloat32Ptr(input string) *float32 {
+	f, err := strconv.ParseFloat(input, 32)
+	if err != nil {
+		return nil
+	}
+	f2 := float32(f)
+	return &f2
+}
+
+func parseToInt32Ptr(input string) *int32 {
+	i, err := strconv.ParseInt(input, 10, 32)
+	if err != nil {
+		return nil
+	}
+	i2 := int32(i)
+	return &i2
+}
+
+func (t *SoundGenerationCMD) Run(ctx *cliContext.Context) error {
+	outputFile := t.OutputFile
+	outputDir := t.BackendAssetsPath
+	if outputFile != "" {
+		outputDir = filepath.Dir(outputFile)
+	}
+
+	text := strings.Join(t.Text, " ")
+
+	externalBackends := make(map[string]string)
+	// split ":" to get backend name and the uri
+	for _, v := range t.ExternalGRPCBackends {
+		backend := v[:strings.IndexByte(v, ':')]
+		uri := v[strings.IndexByte(v, ':')+1:]
+		externalBackends[backend] = uri
+		fmt.Printf("TMP externalBackends[%q]=%q\n\n", backend, uri)
+	}
+
+	opts := &config.ApplicationConfig{
+		ModelPath:            t.ModelsPath,
+		Context:              context.Background(),
+		AudioDir:             outputDir,
+		AssetsDestination:    t.BackendAssetsPath,
+		ExternalGRPCBackends: externalBackends,
+	}
+	ml := model.NewModelLoader(opts.ModelPath)
+
+	defer func() {
+		err := ml.StopAllGRPC()
+		if err != nil {
+			log.Error().Err(err).Msg("unable to stop all grpc processes")
+		}
+	}()
+
+	options := config.BackendConfig{}
+	options.SetDefaults()
+
+	var inputFile *string
+	if t.InputFile != "" {
+		inputFile = &t.InputFile
+	}
+
+	filePath, _, err := backend.SoundGeneration(t.Backend, t.Model, text,
+		parseToFloat32Ptr(t.Duration), parseToFloat32Ptr(t.Temperature), &t.DoSample,
+		inputFile, parseToInt32Ptr(t.InputFileSampleDivisor), ml, opts, options)
+
+	if err != nil {
+		return err
+	}
+	if outputFile != "" {
+		if err := os.Rename(filePath, outputFile); err != nil {
+			return err
+		}
+		fmt.Printf("Generate file %s\n", outputFile)
+	} else {
+		fmt.Printf("Generate file %s\n", filePath)
+	}
+	return nil
+}
diff --git a/core/cli/worker/worker.go b/core/cli/worker/worker.go
index da1b0288..a5d06577 100644
--- a/core/cli/worker/worker.go
+++ b/core/cli/worker/worker.go
@@ -2,6 +2,7 @@ package worker
 
 type WorkerFlags struct {
 	BackendAssetsPath string `env:"LOCALAI_BACKEND_ASSETS_PATH,BACKEND_ASSETS_PATH" type:"path" default:"/tmp/localai/backend_data" help:"Path used to extract libraries that are required by some of the backends in runtime" group:"storage"`
+	ExtraLLamaCPPArgs string `name:"llama-cpp-args" env:"LOCALAI_EXTRA_LLAMA_CPP_ARGS,EXTRA_LLAMA_CPP_ARGS" help:"Extra arguments to pass to llama-cpp-rpc-server"`
 }
 
 type Worker struct {
diff --git a/core/cli/worker/worker_llamacpp.go b/core/cli/worker/worker_llamacpp.go
index 2baf51ec..9fb69ca1 100644
--- a/core/cli/worker/worker_llamacpp.go
+++ b/core/cli/worker/worker_llamacpp.go
@@ -3,6 +3,7 @@ package worker
 import (
 	"fmt"
 	"os"
+	"strings"
 	"syscall"
 
 	cliContext "github.com/mudler/LocalAI/core/cli/context"
@@ -12,7 +13,6 @@ import (
 )
 
 type LLamaCPP struct {
-	Args        []string `arg:"" optional:"" name:"models" help:"Model configuration URLs to load"`
 	WorkerFlags `embed:""`
 }
 
@@ -34,9 +34,8 @@ func (r *LLamaCPP) Run(ctx *cliContext.Context) error {
 		"llama-cpp-rpc-server",
 	)
 
-	args := os.Args[4:]
+	args := strings.Split(r.ExtraLLamaCPPArgs, " ")
 	args, grpcProcess = library.LoadLDSO(r.BackendAssetsPath, args, grpcProcess)
-
 	args = append([]string{grpcProcess}, args...)
 	return syscall.Exec(
 		grpcProcess,
diff --git a/core/cli/worker/worker_p2p.go b/core/cli/worker/worker_p2p.go
index 7c900667..6275481b 100644
--- a/core/cli/worker/worker_p2p.go
+++ b/core/cli/worker/worker_p2p.go
@@ -8,6 +8,7 @@ import (
 	"fmt"
 	"os"
 	"os/exec"
+	"strings"
 	"time"
 
 	cliContext "github.com/mudler/LocalAI/core/cli/context"
@@ -20,12 +21,11 @@ import (
 
 type P2P struct {
 	WorkerFlags        `embed:""`
-	Token              string   `env:"LOCALAI_TOKEN,LOCALAI_P2P_TOKEN,TOKEN" help:"P2P token to use"`
-	NoRunner           bool     `env:"LOCALAI_NO_RUNNER,NO_RUNNER" help:"Do not start the llama-cpp-rpc-server"`
-	RunnerAddress      string   `env:"LOCALAI_RUNNER_ADDRESS,RUNNER_ADDRESS" help:"Address of the llama-cpp-rpc-server"`
-	RunnerPort         string   `env:"LOCALAI_RUNNER_PORT,RUNNER_PORT" help:"Port of the llama-cpp-rpc-server"`
-	ExtraLLamaCPPArgs  []string `env:"LOCALAI_EXTRA_LLAMA_CPP_ARGS,EXTRA_LLAMA_CPP_ARGS" help:"Extra arguments to pass to llama-cpp-rpc-server"`
-	Peer2PeerNetworkID string   `env:"LOCALAI_P2P_NETWORK_ID,P2P_NETWORK_ID" help:"Network ID for P2P mode, can be set arbitrarly by the user for grouping a set of instances" group:"p2p"`
+	Token              string `env:"LOCALAI_TOKEN,LOCALAI_P2P_TOKEN,TOKEN" help:"P2P token to use"`
+	NoRunner           bool   `env:"LOCALAI_NO_RUNNER,NO_RUNNER" help:"Do not start the llama-cpp-rpc-server"`
+	RunnerAddress      string `env:"LOCALAI_RUNNER_ADDRESS,RUNNER_ADDRESS" help:"Address of the llama-cpp-rpc-server"`
+	RunnerPort         string `env:"LOCALAI_RUNNER_PORT,RUNNER_PORT" help:"Port of the llama-cpp-rpc-server"`
+	Peer2PeerNetworkID string `env:"LOCALAI_P2P_NETWORK_ID,P2P_NETWORK_ID" help:"Network ID for P2P mode, can be set arbitrarly by the user for grouping a set of instances" group:"p2p"`
 }
 
 func (r *P2P) Run(ctx *cliContext.Context) error {
@@ -65,44 +65,42 @@ func (r *P2P) Run(ctx *cliContext.Context) error {
 			return err
 		}
 		log.Info().Msgf("You need to start llama-cpp-rpc-server on '%s:%s'", address, p)
+	} else {
+		// Start llama.cpp directly from the version we have pre-packaged
+		go func() {
+			for {
+				log.Info().Msgf("Starting llama-cpp-rpc-server on '%s:%d'", address, port)
 
-		return nil
-	}
+				grpcProcess := assets.ResolvePath(
+					r.BackendAssetsPath,
+					"util",
+					"llama-cpp-rpc-server",
+				)
+				extraArgs := strings.Split(r.ExtraLLamaCPPArgs, " ")
+				args := append([]string{"--host", address, "--port", fmt.Sprint(port)}, extraArgs...)
+				args, grpcProcess = library.LoadLDSO(r.BackendAssetsPath, args, grpcProcess)
 
-	// Start llama.cpp directly from the version we have pre-packaged
-	go func() {
-		for {
-			log.Info().Msgf("Starting llama-cpp-rpc-server on '%s:%d'", address, port)
+				cmd := exec.Command(
+					grpcProcess, args...,
+				)
 
-			grpcProcess := assets.ResolvePath(
-				r.BackendAssetsPath,
-				"util",
-				"llama-cpp-rpc-server",
-			)
+				cmd.Env = os.Environ()
 
-			args := append([]string{"--host", address, "--port", fmt.Sprint(port)}, r.ExtraLLamaCPPArgs...)
-			args, grpcProcess = library.LoadLDSO(r.BackendAssetsPath, args, grpcProcess)
+				cmd.Stderr = os.Stdout
+				cmd.Stdout = os.Stdout
 
-			cmd := exec.Command(
-				grpcProcess, args...,
-			)
+				if err := cmd.Start(); err != nil {
+					log.Error().Any("grpcProcess", grpcProcess).Any("args", args).Err(err).Msg("Failed to start llama-cpp-rpc-server")
+				}
 
-			cmd.Env = os.Environ()
-
-			cmd.Stderr = os.Stdout
-			cmd.Stdout = os.Stdout
-
-			if err := cmd.Start(); err != nil {
-				log.Error().Any("grpcProcess", grpcProcess).Any("args", args).Err(err).Msg("Failed to start llama-cpp-rpc-server")
+				cmd.Wait()
 			}
+		}()
 
-			cmd.Wait()
+		_, err = p2p.ExposeService(context.Background(), address, fmt.Sprint(port), r.Token, p2p.NetworkID(r.Peer2PeerNetworkID, p2p.WorkerID))
+		if err != nil {
+			return err
 		}
-	}()
-
-	_, err = p2p.ExposeService(context.Background(), address, fmt.Sprint(port), r.Token, p2p.NetworkID(r.Peer2PeerNetworkID, p2p.WorkerID))
-	if err != nil {
-		return err
 	}
 
 	for {
diff --git a/core/config/application_config.go b/core/config/application_config.go
index 947c4f13..afbf325f 100644
--- a/core/config/application_config.go
+++ b/core/config/application_config.go
@@ -4,6 +4,7 @@ import (
 	"context"
 	"embed"
 	"encoding/json"
+	"regexp"
 	"time"
 
 	"github.com/mudler/LocalAI/pkg/xsysinfo"
@@ -16,7 +17,6 @@ type ApplicationConfig struct {
 	ModelPath                           string
 	LibPath                             string
 	UploadLimitMB, Threads, ContextSize int
-	DisableWebUI                        bool
 	F16                                 bool
 	Debug                               bool
 	ImageDir                            string
@@ -31,11 +31,17 @@ type ApplicationConfig struct {
 	PreloadModelsFromPath               string
 	CORSAllowOrigins                    string
 	ApiKeys                             []string
-	EnforcePredownloadScans             bool
-	OpaqueErrors                        bool
 	P2PToken                            string
 	P2PNetworkID                        string
 
+	DisableWebUI                       bool
+	EnforcePredownloadScans            bool
+	OpaqueErrors                       bool
+	UseSubtleKeyComparison             bool
+	DisableApiKeyRequirementForHttpGet bool
+	HttpGetExemptedEndpoints           []*regexp.Regexp
+	DisableGalleryEndpoint             bool
+
 	ModelLibraryURL string
 
 	Galleries []Gallery
@@ -57,8 +63,6 @@ type ApplicationConfig struct {
 	ModelsURL []string
 
 	WatchDogBusyTimeout, WatchDogIdleTimeout time.Duration
-
-	DisableGalleryEndpoint bool
 }
 
 type AppOption func(*ApplicationConfig)
@@ -327,6 +331,32 @@ func WithOpaqueErrors(opaque bool) AppOption {
 	}
 }
 
+func WithSubtleKeyComparison(subtle bool) AppOption {
+	return func(o *ApplicationConfig) {
+		o.UseSubtleKeyComparison = subtle
+	}
+}
+
+func WithDisableApiKeyRequirementForHttpGet(required bool) AppOption {
+	return func(o *ApplicationConfig) {
+		o.DisableApiKeyRequirementForHttpGet = required
+	}
+}
+
+func WithHttpGetExemptedEndpoints(endpoints []string) AppOption {
+	return func(o *ApplicationConfig) {
+		o.HttpGetExemptedEndpoints = []*regexp.Regexp{}
+		for _, epr := range endpoints {
+			r, err := regexp.Compile(epr)
+			if err == nil && r != nil {
+				o.HttpGetExemptedEndpoints = append(o.HttpGetExemptedEndpoints, r)
+			} else {
+				log.Warn().Err(err).Str("regex", epr).Msg("Error while compiling HTTP Get Exemption regex, skipping this entry.")
+			}
+		}
+	}
+}
+
 // ToConfigLoaderOptions returns a slice of ConfigLoader Option.
 // Some options defined at the application level are going to be passed as defaults for
 // all the configuration for the models.
diff --git a/core/config/backend_config.go b/core/config/backend_config.go
index ab6a6cc6..5662f1ca 100644
--- a/core/config/backend_config.go
+++ b/core/config/backend_config.go
@@ -126,6 +126,7 @@ type LLMConfig struct {
 	Grammar         string   `yaml:"grammar"`
 	StopWords       []string `yaml:"stopwords"`
 	Cutstrings      []string `yaml:"cutstrings"`
+	ExtractRegex    []string `yaml:"extract_regex"`
 	TrimSpace       []string `yaml:"trimspace"`
 	TrimSuffix      []string `yaml:"trimsuffix"`
 
diff --git a/core/http/app.go b/core/http/app.go
index 6eb9c956..fa9cd866 100644
--- a/core/http/app.go
+++ b/core/http/app.go
@@ -3,13 +3,15 @@ package http
 import (
 	"embed"
 	"errors"
+	"fmt"
 	"net/http"
-	"strings"
 
+	"github.com/dave-gray101/v2keyauth"
 	"github.com/mudler/LocalAI/pkg/utils"
 
 	"github.com/mudler/LocalAI/core/http/endpoints/localai"
 	"github.com/mudler/LocalAI/core/http/endpoints/openai"
+	"github.com/mudler/LocalAI/core/http/middleware"
 	"github.com/mudler/LocalAI/core/http/routes"
 
 	"github.com/mudler/LocalAI/core/config"
@@ -137,37 +139,14 @@ func App(cl *config.BackendConfigLoader, ml *model.ModelLoader, appConfig *confi
 		})
 	}
 
-	// Auth middleware checking if API key is valid. If no API key is set, no auth is required.
-	auth := func(c *fiber.Ctx) error {
-		if len(appConfig.ApiKeys) == 0 {
-			return c.Next()
-		}
-
-		if len(appConfig.ApiKeys) == 0 {
-			return c.Next()
-		}
-
-		authHeader := readAuthHeader(c)
-		if authHeader == "" {
-			return c.Status(fiber.StatusUnauthorized).JSON(fiber.Map{"message": "Authorization header missing"})
-		}
-
-		// If it's a bearer token
-		authHeaderParts := strings.Split(authHeader, " ")
-		if len(authHeaderParts) != 2 || authHeaderParts[0] != "Bearer" {
-			return c.Status(fiber.StatusUnauthorized).JSON(fiber.Map{"message": "Invalid Authorization header format"})
-		}
-
-		apiKey := authHeaderParts[1]
-		for _, key := range appConfig.ApiKeys {
-			if apiKey == key {
-				return c.Next()
-			}
-		}
-
-		return c.Status(fiber.StatusUnauthorized).JSON(fiber.Map{"message": "Invalid API key"})
+	kaConfig, err := middleware.GetKeyAuthConfig(appConfig)
+	if err != nil || kaConfig == nil {
+		return nil, fmt.Errorf("failed to create key auth config: %w", err)
 	}
 
+	// Auth is applied to _all_ endpoints. No exceptions. Filtering out endpoints to bypass is the role of the Filter property of the KeyAuth Configuration
+	app.Use(v2keyauth.New(*kaConfig))
+
 	if appConfig.CORS {
 		var c func(ctx *fiber.Ctx) error
 		if appConfig.CORSAllowOrigins == "" {
@@ -192,13 +171,13 @@ func App(cl *config.BackendConfigLoader, ml *model.ModelLoader, appConfig *confi
 	galleryService := services.NewGalleryService(appConfig)
 	galleryService.Start(appConfig.Context, cl)
 
-	routes.RegisterElevenLabsRoutes(app, cl, ml, appConfig, auth)
-	routes.RegisterLocalAIRoutes(app, cl, ml, appConfig, galleryService, auth)
-	routes.RegisterOpenAIRoutes(app, cl, ml, appConfig, auth)
+	routes.RegisterElevenLabsRoutes(app, cl, ml, appConfig)
+	routes.RegisterLocalAIRoutes(app, cl, ml, appConfig, galleryService)
+	routes.RegisterOpenAIRoutes(app, cl, ml, appConfig)
 	if !appConfig.DisableWebUI {
-		routes.RegisterUIRoutes(app, cl, ml, appConfig, galleryService, auth)
+		routes.RegisterUIRoutes(app, cl, ml, appConfig, galleryService)
 	}
-	routes.RegisterJINARoutes(app, cl, ml, appConfig, auth)
+	routes.RegisterJINARoutes(app, cl, ml, appConfig)
 
 	httpFS := http.FS(embedDirStatic)
 
diff --git a/core/http/app_test.go b/core/http/app_test.go
index a837e20c..86fe7fdd 100644
--- a/core/http/app_test.go
+++ b/core/http/app_test.go
@@ -772,6 +772,17 @@ var _ = Describe("API test", func() {
 			Expect(err.Error()).To(ContainSubstring("error, status code: 500, message: could not load model - all backends returned error:"))
 		})
 
+		It("shows the external backend", func() {
+			// do an http request to the /system endpoint
+			resp, err := http.Get("http://127.0.0.1:9090/system")
+			Expect(err).ToNot(HaveOccurred())
+			Expect(resp.StatusCode).To(Equal(200))
+			dat, err := io.ReadAll(resp.Body)
+			Expect(err).ToNot(HaveOccurred())
+			Expect(string(dat)).To(ContainSubstring("huggingface"))
+			Expect(string(dat)).To(ContainSubstring("llama-cpp"))
+		})
+
 		It("transcribes audio", func() {
 			if runtime.GOOS != "linux" {
 				Skip("test supported only on linux")
diff --git a/core/http/endpoints/elevenlabs/soundgeneration.go b/core/http/endpoints/elevenlabs/soundgeneration.go
new file mode 100644
index 00000000..619544d8
--- /dev/null
+++ b/core/http/endpoints/elevenlabs/soundgeneration.go
@@ -0,0 +1,65 @@
+package elevenlabs
+
+import (
+	"github.com/gofiber/fiber/v2"
+	"github.com/mudler/LocalAI/core/backend"
+	"github.com/mudler/LocalAI/core/config"
+	fiberContext "github.com/mudler/LocalAI/core/http/ctx"
+	"github.com/mudler/LocalAI/core/schema"
+	"github.com/mudler/LocalAI/pkg/model"
+	"github.com/rs/zerolog/log"
+)
+
+// SoundGenerationEndpoint is the ElevenLabs SoundGeneration endpoint https://elevenlabs.io/docs/api-reference/sound-generation
+// @Summary Generates audio from the input text.
+// @Param request body schema.ElevenLabsSoundGenerationRequest true "query params"
+// @Success 200 {string} binary	 "Response"
+// @Router /v1/sound-generation [post]
+func SoundGenerationEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, appConfig *config.ApplicationConfig) func(c *fiber.Ctx) error {
+	return func(c *fiber.Ctx) error {
+		input := new(schema.ElevenLabsSoundGenerationRequest)
+		// Get input data from the request body
+		if err := c.BodyParser(input); err != nil {
+			return err
+		}
+
+		modelFile, err := fiberContext.ModelFromContext(c, cl, ml, input.ModelID, false)
+		if err != nil {
+			modelFile = input.ModelID
+			log.Warn().Str("ModelID", input.ModelID).Msg("Model not found in context")
+		}
+
+		cfg, err := cl.LoadBackendConfigFileByName(modelFile, appConfig.ModelPath,
+			config.LoadOptionDebug(appConfig.Debug),
+			config.LoadOptionThreads(appConfig.Threads),
+			config.LoadOptionContextSize(appConfig.ContextSize),
+			config.LoadOptionF16(appConfig.F16),
+		)
+		if err != nil {
+			modelFile = input.ModelID
+			log.Warn().Str("Request ModelID", input.ModelID).Err(err).Msg("error during LoadBackendConfigFileByName, using request ModelID")
+		} else {
+			if input.ModelID != "" {
+				modelFile = input.ModelID
+			} else {
+				modelFile = cfg.Model
+			}
+		}
+		log.Debug().Str("modelFile", "modelFile").Str("backend", cfg.Backend).Msg("Sound Generation Request about to be sent to backend")
+
+		if input.Duration != nil {
+			log.Debug().Float32("duration", *input.Duration).Msg("duration set")
+		}
+		if input.Temperature != nil {
+			log.Debug().Float32("temperature", *input.Temperature).Msg("temperature set")
+		}
+
+		// TODO: Support uploading files?
+		filePath, _, err := backend.SoundGeneration(cfg.Backend, modelFile, input.Text, input.Duration, input.Temperature, input.DoSample, nil, nil, ml, appConfig, *cfg)
+		if err != nil {
+			return err
+		}
+		return c.Download(filePath)
+
+	}
+}
diff --git a/core/http/endpoints/localai/system.go b/core/http/endpoints/localai/system.go
new file mode 100644
index 00000000..11704933
--- /dev/null
+++ b/core/http/endpoints/localai/system.go
@@ -0,0 +1,29 @@
+package localai
+
+import (
+	"github.com/gofiber/fiber/v2"
+	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/schema"
+	"github.com/mudler/LocalAI/pkg/model"
+)
+
+// SystemInformations returns the system informations
+// @Summary Show the LocalAI instance information
+// @Success 200 {object} schema.SystemInformationResponse "Response"
+// @Router /system [get]
+func SystemInformations(ml *model.ModelLoader, appConfig *config.ApplicationConfig) func(*fiber.Ctx) error {
+	return func(c *fiber.Ctx) error {
+		availableBackends, err := ml.ListAvailableBackends(appConfig.AssetsDestination)
+		if err != nil {
+			return err
+		}
+		for b := range appConfig.ExternalGRPCBackends {
+			availableBackends = append(availableBackends, b)
+		}
+		return c.JSON(
+			schema.SystemInformationResponse{
+				Backends: availableBackends,
+			},
+		)
+	}
+}
diff --git a/core/http/endpoints/openai/chat.go b/core/http/endpoints/openai/chat.go
index 12a14eac..8144bdcd 100644
--- a/core/http/endpoints/openai/chat.go
+++ b/core/http/endpoints/openai/chat.go
@@ -25,9 +25,8 @@ import (
 // @Success 200 {object} schema.OpenAIResponse "Response"
 // @Router /v1/chat/completions [post]
 func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, startupOptions *config.ApplicationConfig) func(c *fiber.Ctx) error {
-	textContentToReturn := ""
-	id := uuid.New().String()
-	created := int(time.Now().Unix())
+	var id, textContentToReturn string
+	var created int
 
 	process := func(s string, req *schema.OpenAIRequest, config *config.BackendConfig, loader *model.ModelLoader, responses chan schema.OpenAIResponse) {
 		initialMessage := schema.OpenAIResponse{
@@ -69,9 +68,9 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, startup
 
 		textContentToReturn = functions.ParseTextContent(result, config.FunctionsConfig)
 		result = functions.CleanupLLMResult(result, config.FunctionsConfig)
-		results := functions.ParseFunctionCall(result, config.FunctionsConfig)
+		functionResults := functions.ParseFunctionCall(result, config.FunctionsConfig)
 		log.Debug().Msgf("Text content to return: %s", textContentToReturn)
-		noActionToRun := len(results) > 0 && results[0].Name == noAction || len(results) == 0
+		noActionToRun := len(functionResults) > 0 && functionResults[0].Name == noAction || len(functionResults) == 0
 
 		switch {
 		case noActionToRun:
@@ -84,7 +83,7 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, startup
 			}
 			responses <- initialMessage
 
-			result, err := handleQuestion(config, req, ml, startupOptions, results, result, prompt)
+			result, err := handleQuestion(config, req, ml, startupOptions, functionResults, result, prompt)
 			if err != nil {
 				log.Error().Err(err).Msg("error handling question")
 				return
@@ -106,7 +105,7 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, startup
 			responses <- resp
 
 		default:
-			for i, ss := range results {
+			for i, ss := range functionResults {
 				name, args := ss.Name, ss.Arguments
 
 				initialMessage := schema.OpenAIResponse{
@@ -159,6 +158,10 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, startup
 	}
 
 	return func(c *fiber.Ctx) error {
+		textContentToReturn = ""
+		id = uuid.New().String()
+		created = int(time.Now().Unix())
+
 		modelFile, input, err := readRequest(c, cl, ml, startupOptions, true)
 		if err != nil {
 			return fmt.Errorf("failed reading parameters from request:%w", err)
diff --git a/core/http/middleware/auth.go b/core/http/middleware/auth.go
new file mode 100644
index 00000000..bc8bcf80
--- /dev/null
+++ b/core/http/middleware/auth.go
@@ -0,0 +1,93 @@
+package middleware
+
+import (
+	"crypto/subtle"
+	"errors"
+
+	"github.com/dave-gray101/v2keyauth"
+	"github.com/gofiber/fiber/v2"
+	"github.com/gofiber/fiber/v2/middleware/keyauth"
+	"github.com/mudler/LocalAI/core/config"
+)
+
+// This file contains the configuration generators and handler functions that are used along with the fiber/keyauth middleware
+// Currently this requires an upstream patch - and feature patches are no longer accepted to v2
+// Therefore `dave-gray101/v2keyauth` contains the v2 backport of the middleware until v3 stabilizes and we migrate.
+
+func GetKeyAuthConfig(applicationConfig *config.ApplicationConfig) (*v2keyauth.Config, error) {
+	customLookup, err := v2keyauth.MultipleKeySourceLookup([]string{"header:Authorization", "header:x-api-key", "header:xi-api-key"}, keyauth.ConfigDefault.AuthScheme)
+	if err != nil {
+		return nil, err
+	}
+
+	return &v2keyauth.Config{
+		CustomKeyLookup: customLookup,
+		Next:            getApiKeyRequiredFilterFunction(applicationConfig),
+		Validator:       getApiKeyValidationFunction(applicationConfig),
+		ErrorHandler:    getApiKeyErrorHandler(applicationConfig),
+		AuthScheme:      "Bearer",
+	}, nil
+}
+
+func getApiKeyErrorHandler(applicationConfig *config.ApplicationConfig) fiber.ErrorHandler {
+	return func(ctx *fiber.Ctx, err error) error {
+		if errors.Is(err, v2keyauth.ErrMissingOrMalformedAPIKey) {
+			if len(applicationConfig.ApiKeys) == 0 {
+				return ctx.Next() // if no keys are set up, any error we get here is not an error.
+			}
+			if applicationConfig.OpaqueErrors {
+				return ctx.SendStatus(403)
+			}
+		}
+		if applicationConfig.OpaqueErrors {
+			return ctx.SendStatus(500)
+		}
+		return err
+	}
+}
+
+func getApiKeyValidationFunction(applicationConfig *config.ApplicationConfig) func(*fiber.Ctx, string) (bool, error) {
+
+	if applicationConfig.UseSubtleKeyComparison {
+		return func(ctx *fiber.Ctx, apiKey string) (bool, error) {
+			if len(applicationConfig.ApiKeys) == 0 {
+				return true, nil // If no keys are setup, accept everything
+			}
+			for _, validKey := range applicationConfig.ApiKeys {
+				if subtle.ConstantTimeCompare([]byte(apiKey), []byte(validKey)) == 1 {
+					return true, nil
+				}
+			}
+			return false, v2keyauth.ErrMissingOrMalformedAPIKey
+		}
+	}
+
+	return func(ctx *fiber.Ctx, apiKey string) (bool, error) {
+		if len(applicationConfig.ApiKeys) == 0 {
+			return true, nil // If no keys are setup, accept everything
+		}
+		for _, validKey := range applicationConfig.ApiKeys {
+			if apiKey == validKey {
+				return true, nil
+			}
+		}
+		return false, v2keyauth.ErrMissingOrMalformedAPIKey
+	}
+}
+
+func getApiKeyRequiredFilterFunction(applicationConfig *config.ApplicationConfig) func(*fiber.Ctx) bool {
+	if applicationConfig.DisableApiKeyRequirementForHttpGet {
+		return func(c *fiber.Ctx) bool {
+			if c.Method() != "GET" {
+				return false
+			}
+			for _, rx := range applicationConfig.HttpGetExemptedEndpoints {
+				if rx.MatchString(c.Path()) {
+					return true
+				}
+			}
+			return false
+		}
+	}
+	return func(c *fiber.Ctx) bool { return false }
+}
\ No newline at end of file
diff --git a/core/http/routes/elevenlabs.go b/core/http/routes/elevenlabs.go
index 4f9e666f..73387c7b 100644
--- a/core/http/routes/elevenlabs.go
+++ b/core/http/routes/elevenlabs.go
@@ -10,10 +10,11 @@ import (
 func RegisterElevenLabsRoutes(app *fiber.App,
 	cl *config.BackendConfigLoader,
 	ml *model.ModelLoader,
-	appConfig *config.ApplicationConfig,
-	auth func(*fiber.Ctx) error) {
+	appConfig *config.ApplicationConfig) {
 
 	// Elevenlabs
-	app.Post("/v1/text-to-speech/:voice-id", auth, elevenlabs.TTSEndpoint(cl, ml, appConfig))
+	app.Post("/v1/text-to-speech/:voice-id", elevenlabs.TTSEndpoint(cl, ml, appConfig))
+
+	app.Post("/v1/sound-generation", elevenlabs.SoundGenerationEndpoint(cl, ml, appConfig))
 
 }
diff --git a/core/http/routes/jina.go b/core/http/routes/jina.go
index 92f29224..93125e6c 100644
--- a/core/http/routes/jina.go
+++ b/core/http/routes/jina.go
@@ -11,8 +11,7 @@ import (
 func RegisterJINARoutes(app *fiber.App,
 	cl *config.BackendConfigLoader,
 	ml *model.ModelLoader,
-	appConfig *config.ApplicationConfig,
-	auth func(*fiber.Ctx) error) {
+	appConfig *config.ApplicationConfig) {
 
 	// POST endpoint to mimic the reranking
 	app.Post("/v1/rerank", jina.JINARerankEndpoint(cl, ml, appConfig))
diff --git a/core/http/routes/localai.go b/core/http/routes/localai.go
index 105991e8..29fef378 100644
--- a/core/http/routes/localai.go
+++ b/core/http/routes/localai.go
@@ -15,33 +15,32 @@ func RegisterLocalAIRoutes(app *fiber.App,
 	cl *config.BackendConfigLoader,
 	ml *model.ModelLoader,
 	appConfig *config.ApplicationConfig,
-	galleryService *services.GalleryService,
-	auth func(*fiber.Ctx) error) {
+	galleryService *services.GalleryService) {
 
 	app.Get("/swagger/*", swagger.HandlerDefault) // default
 
 	// LocalAI API endpoints
 	if !appConfig.DisableGalleryEndpoint {
 		modelGalleryEndpointService := localai.CreateModelGalleryEndpointService(appConfig.Galleries, appConfig.ModelPath, galleryService)
-		app.Post("/models/apply", auth, modelGalleryEndpointService.ApplyModelGalleryEndpoint())
-		app.Post("/models/delete/:name", auth, modelGalleryEndpointService.DeleteModelGalleryEndpoint())
+		app.Post("/models/apply", modelGalleryEndpointService.ApplyModelGalleryEndpoint())
+		app.Post("/models/delete/:name", modelGalleryEndpointService.DeleteModelGalleryEndpoint())
 
-		app.Get("/models/available", auth, modelGalleryEndpointService.ListModelFromGalleryEndpoint())
-		app.Get("/models/galleries", auth, modelGalleryEndpointService.ListModelGalleriesEndpoint())
-		app.Post("/models/galleries", auth, modelGalleryEndpointService.AddModelGalleryEndpoint())
-		app.Delete("/models/galleries", auth, modelGalleryEndpointService.RemoveModelGalleryEndpoint())
-		app.Get("/models/jobs/:uuid", auth, modelGalleryEndpointService.GetOpStatusEndpoint())
-		app.Get("/models/jobs", auth, modelGalleryEndpointService.GetAllStatusEndpoint())
+		app.Get("/models/available", modelGalleryEndpointService.ListModelFromGalleryEndpoint())
+		app.Get("/models/galleries", modelGalleryEndpointService.ListModelGalleriesEndpoint())
+		app.Post("/models/galleries", modelGalleryEndpointService.AddModelGalleryEndpoint())
+		app.Delete("/models/galleries", modelGalleryEndpointService.RemoveModelGalleryEndpoint())
+		app.Get("/models/jobs/:uuid", modelGalleryEndpointService.GetOpStatusEndpoint())
+		app.Get("/models/jobs", modelGalleryEndpointService.GetAllStatusEndpoint())
 	}
 
-	app.Post("/tts", auth, localai.TTSEndpoint(cl, ml, appConfig))
+	app.Post("/tts", localai.TTSEndpoint(cl, ml, appConfig))
 
 	// Stores
 	sl := model.NewModelLoader("")
-	app.Post("/stores/set", auth, localai.StoresSetEndpoint(sl, appConfig))
-	app.Post("/stores/delete", auth, localai.StoresDeleteEndpoint(sl, appConfig))
-	app.Post("/stores/get", auth, localai.StoresGetEndpoint(sl, appConfig))
-	app.Post("/stores/find", auth, localai.StoresFindEndpoint(sl, appConfig))
+	app.Post("/stores/set", localai.StoresSetEndpoint(sl, appConfig))
+	app.Post("/stores/delete", localai.StoresDeleteEndpoint(sl, appConfig))
+	app.Post("/stores/get", localai.StoresGetEndpoint(sl, appConfig))
+	app.Post("/stores/find", localai.StoresFindEndpoint(sl, appConfig))
 
 	// Kubernetes health checks
 	ok := func(c *fiber.Ctx) error {
@@ -51,23 +50,25 @@ func RegisterLocalAIRoutes(app *fiber.App,
 	app.Get("/healthz", ok)
 	app.Get("/readyz", ok)
 
-	app.Get("/metrics", auth, localai.LocalAIMetricsEndpoint())
+	app.Get("/metrics", localai.LocalAIMetricsEndpoint())
 
 	// Experimental Backend Statistics Module
 	backendMonitorService := services.NewBackendMonitorService(ml, cl, appConfig) // Split out for now
-	app.Get("/backend/monitor", auth, localai.BackendMonitorEndpoint(backendMonitorService))
-	app.Post("/backend/shutdown", auth, localai.BackendShutdownEndpoint(backendMonitorService))
+	app.Get("/backend/monitor", localai.BackendMonitorEndpoint(backendMonitorService))
+	app.Post("/backend/shutdown", localai.BackendShutdownEndpoint(backendMonitorService))
 
 	// p2p
 	if p2p.IsP2PEnabled() {
-		app.Get("/api/p2p", auth, localai.ShowP2PNodes(appConfig))
-		app.Get("/api/p2p/token", auth, localai.ShowP2PToken(appConfig))
+		app.Get("/api/p2p", localai.ShowP2PNodes(appConfig))
+		app.Get("/api/p2p/token", localai.ShowP2PToken(appConfig))
 	}
 
-	app.Get("/version", auth, func(c *fiber.Ctx) error {
+	app.Get("/version", func(c *fiber.Ctx) error {
 		return c.JSON(struct {
 			Version string `json:"version"`
 		}{Version: internal.PrintableVersion()})
 	})
 
+	app.Get("/system", auth, localai.SystemInformations(ml, appConfig))
+
 }
diff --git a/core/http/routes/openai.go b/core/http/routes/openai.go
index e190bc6d..081daf70 100644
--- a/core/http/routes/openai.go
+++ b/core/http/routes/openai.go
@@ -11,66 +11,65 @@ import (
 func RegisterOpenAIRoutes(app *fiber.App,
 	cl *config.BackendConfigLoader,
 	ml *model.ModelLoader,
-	appConfig *config.ApplicationConfig,
-	auth func(*fiber.Ctx) error) {
+	appConfig *config.ApplicationConfig) {
 	// openAI compatible API endpoint
 
 	// chat
-	app.Post("/v1/chat/completions", auth, openai.ChatEndpoint(cl, ml, appConfig))
-	app.Post("/chat/completions", auth, openai.ChatEndpoint(cl, ml, appConfig))
+	app.Post("/v1/chat/completions", openai.ChatEndpoint(cl, ml, appConfig))
+	app.Post("/chat/completions", openai.ChatEndpoint(cl, ml, appConfig))
 
 	// edit
-	app.Post("/v1/edits", auth, openai.EditEndpoint(cl, ml, appConfig))
-	app.Post("/edits", auth, openai.EditEndpoint(cl, ml, appConfig))
+	app.Post("/v1/edits", openai.EditEndpoint(cl, ml, appConfig))
+	app.Post("/edits", openai.EditEndpoint(cl, ml, appConfig))
 
 	// assistant
-	app.Get("/v1/assistants", auth, openai.ListAssistantsEndpoint(cl, ml, appConfig))
-	app.Get("/assistants", auth, openai.ListAssistantsEndpoint(cl, ml, appConfig))
-	app.Post("/v1/assistants", auth, openai.CreateAssistantEndpoint(cl, ml, appConfig))
-	app.Post("/assistants", auth, openai.CreateAssistantEndpoint(cl, ml, appConfig))
-	app.Delete("/v1/assistants/:assistant_id", auth, openai.DeleteAssistantEndpoint(cl, ml, appConfig))
-	app.Delete("/assistants/:assistant_id", auth, openai.DeleteAssistantEndpoint(cl, ml, appConfig))
-	app.Get("/v1/assistants/:assistant_id", auth, openai.GetAssistantEndpoint(cl, ml, appConfig))
-	app.Get("/assistants/:assistant_id", auth, openai.GetAssistantEndpoint(cl, ml, appConfig))
-	app.Post("/v1/assistants/:assistant_id", auth, openai.ModifyAssistantEndpoint(cl, ml, appConfig))
-	app.Post("/assistants/:assistant_id", auth, openai.ModifyAssistantEndpoint(cl, ml, appConfig))
-	app.Get("/v1/assistants/:assistant_id/files", auth, openai.ListAssistantFilesEndpoint(cl, ml, appConfig))
-	app.Get("/assistants/:assistant_id/files", auth, openai.ListAssistantFilesEndpoint(cl, ml, appConfig))
-	app.Post("/v1/assistants/:assistant_id/files", auth, openai.CreateAssistantFileEndpoint(cl, ml, appConfig))
-	app.Post("/assistants/:assistant_id/files", auth, openai.CreateAssistantFileEndpoint(cl, ml, appConfig))
-	app.Delete("/v1/assistants/:assistant_id/files/:file_id", auth, openai.DeleteAssistantFileEndpoint(cl, ml, appConfig))
-	app.Delete("/assistants/:assistant_id/files/:file_id", auth, openai.DeleteAssistantFileEndpoint(cl, ml, appConfig))
-	app.Get("/v1/assistants/:assistant_id/files/:file_id", auth, openai.GetAssistantFileEndpoint(cl, ml, appConfig))
-	app.Get("/assistants/:assistant_id/files/:file_id", auth, openai.GetAssistantFileEndpoint(cl, ml, appConfig))
+	app.Get("/v1/assistants", openai.ListAssistantsEndpoint(cl, ml, appConfig))
+	app.Get("/assistants", openai.ListAssistantsEndpoint(cl, ml, appConfig))
+	app.Post("/v1/assistants", openai.CreateAssistantEndpoint(cl, ml, appConfig))
+	app.Post("/assistants", openai.CreateAssistantEndpoint(cl, ml, appConfig))
+	app.Delete("/v1/assistants/:assistant_id", openai.DeleteAssistantEndpoint(cl, ml, appConfig))
+	app.Delete("/assistants/:assistant_id", openai.DeleteAssistantEndpoint(cl, ml, appConfig))
+	app.Get("/v1/assistants/:assistant_id", openai.GetAssistantEndpoint(cl, ml, appConfig))
+	app.Get("/assistants/:assistant_id", openai.GetAssistantEndpoint(cl, ml, appConfig))
+	app.Post("/v1/assistants/:assistant_id", openai.ModifyAssistantEndpoint(cl, ml, appConfig))
+	app.Post("/assistants/:assistant_id", openai.ModifyAssistantEndpoint(cl, ml, appConfig))
+	app.Get("/v1/assistants/:assistant_id/files", openai.ListAssistantFilesEndpoint(cl, ml, appConfig))
+	app.Get("/assistants/:assistant_id/files", openai.ListAssistantFilesEndpoint(cl, ml, appConfig))
+	app.Post("/v1/assistants/:assistant_id/files", openai.CreateAssistantFileEndpoint(cl, ml, appConfig))
+	app.Post("/assistants/:assistant_id/files", openai.CreateAssistantFileEndpoint(cl, ml, appConfig))
+	app.Delete("/v1/assistants/:assistant_id/files/:file_id", openai.DeleteAssistantFileEndpoint(cl, ml, appConfig))
+	app.Delete("/assistants/:assistant_id/files/:file_id", openai.DeleteAssistantFileEndpoint(cl, ml, appConfig))
+	app.Get("/v1/assistants/:assistant_id/files/:file_id", openai.GetAssistantFileEndpoint(cl, ml, appConfig))
+	app.Get("/assistants/:assistant_id/files/:file_id", openai.GetAssistantFileEndpoint(cl, ml, appConfig))
 
 	// files
-	app.Post("/v1/files", auth, openai.UploadFilesEndpoint(cl, appConfig))
-	app.Post("/files", auth, openai.UploadFilesEndpoint(cl, appConfig))
-	app.Get("/v1/files", auth, openai.ListFilesEndpoint(cl, appConfig))
-	app.Get("/files", auth, openai.ListFilesEndpoint(cl, appConfig))
-	app.Get("/v1/files/:file_id", auth, openai.GetFilesEndpoint(cl, appConfig))
-	app.Get("/files/:file_id", auth, openai.GetFilesEndpoint(cl, appConfig))
-	app.Delete("/v1/files/:file_id", auth, openai.DeleteFilesEndpoint(cl, appConfig))
-	app.Delete("/files/:file_id", auth, openai.DeleteFilesEndpoint(cl, appConfig))
-	app.Get("/v1/files/:file_id/content", auth, openai.GetFilesContentsEndpoint(cl, appConfig))
-	app.Get("/files/:file_id/content", auth, openai.GetFilesContentsEndpoint(cl, appConfig))
+	app.Post("/v1/files", openai.UploadFilesEndpoint(cl, appConfig))
+	app.Post("/files", openai.UploadFilesEndpoint(cl, appConfig))
+	app.Get("/v1/files", openai.ListFilesEndpoint(cl, appConfig))
+	app.Get("/files", openai.ListFilesEndpoint(cl, appConfig))
+	app.Get("/v1/files/:file_id", openai.GetFilesEndpoint(cl, appConfig))
+	app.Get("/files/:file_id", openai.GetFilesEndpoint(cl, appConfig))
+	app.Delete("/v1/files/:file_id", openai.DeleteFilesEndpoint(cl, appConfig))
+	app.Delete("/files/:file_id", openai.DeleteFilesEndpoint(cl, appConfig))
+	app.Get("/v1/files/:file_id/content", openai.GetFilesContentsEndpoint(cl, appConfig))
+	app.Get("/files/:file_id/content", openai.GetFilesContentsEndpoint(cl, appConfig))
 
 	// completion
-	app.Post("/v1/completions", auth, openai.CompletionEndpoint(cl, ml, appConfig))
-	app.Post("/completions", auth, openai.CompletionEndpoint(cl, ml, appConfig))
-	app.Post("/v1/engines/:model/completions", auth, openai.CompletionEndpoint(cl, ml, appConfig))
+	app.Post("/v1/completions", openai.CompletionEndpoint(cl, ml, appConfig))
+	app.Post("/completions", openai.CompletionEndpoint(cl, ml, appConfig))
+	app.Post("/v1/engines/:model/completions", openai.CompletionEndpoint(cl, ml, appConfig))
 
 	// embeddings
-	app.Post("/v1/embeddings", auth, openai.EmbeddingsEndpoint(cl, ml, appConfig))
-	app.Post("/embeddings", auth, openai.EmbeddingsEndpoint(cl, ml, appConfig))
-	app.Post("/v1/engines/:model/embeddings", auth, openai.EmbeddingsEndpoint(cl, ml, appConfig))
+	app.Post("/v1/embeddings", openai.EmbeddingsEndpoint(cl, ml, appConfig))
+	app.Post("/embeddings", openai.EmbeddingsEndpoint(cl, ml, appConfig))
+	app.Post("/v1/engines/:model/embeddings", openai.EmbeddingsEndpoint(cl, ml, appConfig))
 
 	// audio
-	app.Post("/v1/audio/transcriptions", auth, openai.TranscriptEndpoint(cl, ml, appConfig))
-	app.Post("/v1/audio/speech", auth, localai.TTSEndpoint(cl, ml, appConfig))
+	app.Post("/v1/audio/transcriptions", openai.TranscriptEndpoint(cl, ml, appConfig))
+	app.Post("/v1/audio/speech", localai.TTSEndpoint(cl, ml, appConfig))
 
 	// images
-	app.Post("/v1/images/generations", auth, openai.ImageEndpoint(cl, ml, appConfig))
+	app.Post("/v1/images/generations", openai.ImageEndpoint(cl, ml, appConfig))
 
 	if appConfig.ImageDir != "" {
 		app.Static("/generated-images", appConfig.ImageDir)
@@ -81,6 +80,6 @@ func RegisterOpenAIRoutes(app *fiber.App,
 	}
 
 	// List models
-	app.Get("/v1/models", auth, openai.ListModelsEndpoint(cl, ml))
-	app.Get("/models", auth, openai.ListModelsEndpoint(cl, ml))
+	app.Get("/v1/models", openai.ListModelsEndpoint(cl, ml))
+	app.Get("/models", openai.ListModelsEndpoint(cl, ml))
 }
diff --git a/core/http/routes/ui.go b/core/http/routes/ui.go
index 6dfb3f43..7b2c6ae7 100644
--- a/core/http/routes/ui.go
+++ b/core/http/routes/ui.go
@@ -59,8 +59,7 @@ func RegisterUIRoutes(app *fiber.App,
 	cl *config.BackendConfigLoader,
 	ml *model.ModelLoader,
 	appConfig *config.ApplicationConfig,
-	galleryService *services.GalleryService,
-	auth func(*fiber.Ctx) error) {
+	galleryService *services.GalleryService) {
 
 	// keeps the state of models that are being installed from the UI
 	var processingModels = NewModelOpCache()
@@ -85,10 +84,10 @@ func RegisterUIRoutes(app *fiber.App,
 		return processingModelsData, taskTypes
 	}
 
-	app.Get("/", auth, localai.WelcomeEndpoint(appConfig, cl, ml, modelStatus))
+	app.Get("/", localai.WelcomeEndpoint(appConfig, cl, ml, modelStatus))
 
 	if p2p.IsP2PEnabled() {
-		app.Get("/p2p", auth, func(c *fiber.Ctx) error {
+		app.Get("/p2p", func(c *fiber.Ctx) error {
 			summary := fiber.Map{
 				"Title":   "LocalAI - P2P dashboard",
 				"Version": internal.PrintableVersion(),
@@ -104,17 +103,17 @@ func RegisterUIRoutes(app *fiber.App,
 		})
 
 		/* show nodes live! */
-		app.Get("/p2p/ui/workers", auth, func(c *fiber.Ctx) error {
+		app.Get("/p2p/ui/workers", func(c *fiber.Ctx) error {
 			return c.SendString(elements.P2PNodeBoxes(p2p.GetAvailableNodes(p2p.NetworkID(appConfig.P2PNetworkID, p2p.WorkerID))))
 		})
-		app.Get("/p2p/ui/workers-federation", auth, func(c *fiber.Ctx) error {
+		app.Get("/p2p/ui/workers-federation", func(c *fiber.Ctx) error {
 			return c.SendString(elements.P2PNodeBoxes(p2p.GetAvailableNodes(p2p.NetworkID(appConfig.P2PNetworkID, p2p.FederatedID))))
 		})
 
-		app.Get("/p2p/ui/workers-stats", auth, func(c *fiber.Ctx) error {
+		app.Get("/p2p/ui/workers-stats", func(c *fiber.Ctx) error {
 			return c.SendString(elements.P2PNodeStats(p2p.GetAvailableNodes(p2p.NetworkID(appConfig.P2PNetworkID, p2p.WorkerID))))
 		})
-		app.Get("/p2p/ui/workers-federation-stats", auth, func(c *fiber.Ctx) error {
+		app.Get("/p2p/ui/workers-federation-stats", func(c *fiber.Ctx) error {
 			return c.SendString(elements.P2PNodeStats(p2p.GetAvailableNodes(p2p.NetworkID(appConfig.P2PNetworkID, p2p.FederatedID))))
 		})
 	}
@@ -122,7 +121,7 @@ func RegisterUIRoutes(app *fiber.App,
 	if !appConfig.DisableGalleryEndpoint {
 
 		// Show the Models page (all models)
-		app.Get("/browse", auth, func(c *fiber.Ctx) error {
+		app.Get("/browse", func(c *fiber.Ctx) error {
 			term := c.Query("term")
 
 			models, _ := gallery.AvailableGalleryModels(appConfig.Galleries, appConfig.ModelPath)
@@ -167,7 +166,7 @@ func RegisterUIRoutes(app *fiber.App,
 
 		// Show the models, filtered from the user input
 		// https://htmx.org/examples/active-search/
-		app.Post("/browse/search/models", auth, func(c *fiber.Ctx) error {
+		app.Post("/browse/search/models", func(c *fiber.Ctx) error {
 			form := struct {
 				Search string `form:"search"`
 			}{}
@@ -188,7 +187,7 @@ func RegisterUIRoutes(app *fiber.App,
 
 		// This route is used when the "Install" button is pressed, we submit here a new job to the gallery service
 		// https://htmx.org/examples/progress-bar/
-		app.Post("/browse/install/model/:id", auth, func(c *fiber.Ctx) error {
+		app.Post("/browse/install/model/:id", func(c *fiber.Ctx) error {
 			galleryID := strings.Clone(c.Params("id")) // note: strings.Clone is required for multiple requests!
 			log.Debug().Msgf("UI job submitted to install  : %+v\n", galleryID)
 
@@ -215,7 +214,7 @@ func RegisterUIRoutes(app *fiber.App,
 
 		// This route is used when the "Install" button is pressed, we submit here a new job to the gallery service
 		// https://htmx.org/examples/progress-bar/
-		app.Post("/browse/delete/model/:id", auth, func(c *fiber.Ctx) error {
+		app.Post("/browse/delete/model/:id", func(c *fiber.Ctx) error {
 			galleryID := strings.Clone(c.Params("id")) // note: strings.Clone is required for multiple requests!
 			log.Debug().Msgf("UI job submitted to delete  : %+v\n", galleryID)
 			var galleryName = galleryID
@@ -255,7 +254,7 @@ func RegisterUIRoutes(app *fiber.App,
 		// Display the job current progress status
 		// If the job is done, we trigger the /browse/job/:uid route
 		// https://htmx.org/examples/progress-bar/
-		app.Get("/browse/job/progress/:uid", auth, func(c *fiber.Ctx) error {
+		app.Get("/browse/job/progress/:uid", func(c *fiber.Ctx) error {
 			jobUID := strings.Clone(c.Params("uid")) // note: strings.Clone is required for multiple requests!
 
 			status := galleryService.GetStatus(jobUID)
@@ -279,7 +278,7 @@ func RegisterUIRoutes(app *fiber.App,
 
 		// this route is hit when the job is done, and we display the
 		// final state (for now just displays "Installation completed")
-		app.Get("/browse/job/:uid", auth, func(c *fiber.Ctx) error {
+		app.Get("/browse/job/:uid", func(c *fiber.Ctx) error {
 			jobUID := strings.Clone(c.Params("uid")) // note: strings.Clone is required for multiple requests!
 
 			status := galleryService.GetStatus(jobUID)
@@ -303,7 +302,7 @@ func RegisterUIRoutes(app *fiber.App,
 	}
 
 	// Show the Chat page
-	app.Get("/chat/:model", auth, func(c *fiber.Ctx) error {
+	app.Get("/chat/:model", func(c *fiber.Ctx) error {
 		backendConfigs, _ := services.ListModels(cl, ml, "", true)
 
 		summary := fiber.Map{
@@ -318,7 +317,7 @@ func RegisterUIRoutes(app *fiber.App,
 		return c.Render("views/chat", summary)
 	})
 
-	app.Get("/talk/", auth, func(c *fiber.Ctx) error {
+	app.Get("/talk/", func(c *fiber.Ctx) error {
 		backendConfigs, _ := services.ListModels(cl, ml, "", true)
 
 		if len(backendConfigs) == 0 {
@@ -338,7 +337,7 @@ func RegisterUIRoutes(app *fiber.App,
 		return c.Render("views/talk", summary)
 	})
 
-	app.Get("/chat/", auth, func(c *fiber.Ctx) error {
+	app.Get("/chat/", func(c *fiber.Ctx) error {
 
 		backendConfigs, _ := services.ListModels(cl, ml, "", true)
 
@@ -359,7 +358,7 @@ func RegisterUIRoutes(app *fiber.App,
 		return c.Render("views/chat", summary)
 	})
 
-	app.Get("/text2image/:model", auth, func(c *fiber.Ctx) error {
+	app.Get("/text2image/:model", func(c *fiber.Ctx) error {
 		backendConfigs := cl.GetAllBackendConfigs()
 
 		summary := fiber.Map{
@@ -374,7 +373,7 @@ func RegisterUIRoutes(app *fiber.App,
 		return c.Render("views/text2image", summary)
 	})
 
-	app.Get("/text2image/", auth, func(c *fiber.Ctx) error {
+	app.Get("/text2image/", func(c *fiber.Ctx) error {
 
 		backendConfigs := cl.GetAllBackendConfigs()
 
@@ -395,7 +394,7 @@ func RegisterUIRoutes(app *fiber.App,
 		return c.Render("views/text2image", summary)
 	})
 
-	app.Get("/tts/:model", auth, func(c *fiber.Ctx) error {
+	app.Get("/tts/:model", func(c *fiber.Ctx) error {
 		backendConfigs := cl.GetAllBackendConfigs()
 
 		summary := fiber.Map{
@@ -410,7 +409,7 @@ func RegisterUIRoutes(app *fiber.App,
 		return c.Render("views/tts", summary)
 	})
 
-	app.Get("/tts/", auth, func(c *fiber.Ctx) error {
+	app.Get("/tts/", func(c *fiber.Ctx) error {
 
 		backendConfigs := cl.GetAllBackendConfigs()
 
diff --git a/core/http/views/partials/head.html b/core/http/views/partials/head.html
index 8d072093..5c119fba 100644
--- a/core/http/views/partials/head.html
+++ b/core/http/views/partials/head.html
@@ -6,11 +6,7 @@
     rel="stylesheet"
     href="/static/assets/highlightjs.css"
   />
-  <script defer src="/static/assets/anime.min.js"></script>
-    <script
-    defer
-    src="/static/assets/highlightjs.js"
-  ></script>
+  <script defer src="/static/assets/highlightjs.js"></script>
     <script
     defer
     src="/static/assets/alpine.js"
diff --git a/core/p2p/p2p.go b/core/p2p/p2p.go
index fe55346a..15e1dc37 100644
--- a/core/p2p/p2p.go
+++ b/core/p2p/p2p.go
@@ -28,9 +28,15 @@ import (
 	"github.com/mudler/edgevpn/pkg/logger"
 )
 
-func generateNewConnectionData() *node.YAMLConnectionConfig {
+func generateNewConnectionData(DHTInterval, OTPInterval int) *node.YAMLConnectionConfig {
 	maxMessSize := 20 << 20 // 20MB
 	keyLength := 43
+	if DHTInterval == 0 {
+		DHTInterval = 360
+	}
+	if OTPInterval == 0 {
+		OTPInterval = 9000
+	}
 
 	return &node.YAMLConnectionConfig{
 		MaxMessageSize: maxMessSize,
@@ -40,21 +46,21 @@ func generateNewConnectionData() *node.YAMLConnectionConfig {
 		OTP: node.OTP{
 			DHT: node.OTPConfig{
 				Key:      eutils.RandStringRunes(keyLength),
-				Interval: 120,
+				Interval: DHTInterval,
 				Length:   keyLength,
 			},
 			Crypto: node.OTPConfig{
 				Key:      eutils.RandStringRunes(keyLength),
-				Interval: 9000,
+				Interval: OTPInterval,
 				Length:   keyLength,
 			},
 		},
 	}
 }
 
-func GenerateToken() string {
+func GenerateToken(DHTInterval, OTPInterval int) string {
 	// Generates a new config and exit
-	return generateNewConnectionData().Base64()
+	return generateNewConnectionData(DHTInterval, OTPInterval).Base64()
 }
 
 func IsP2PEnabled() bool {
@@ -202,13 +208,9 @@ func ServiceDiscoverer(ctx context.Context, n *node.Node, token, servicesID stri
 func discoveryTunnels(ctx context.Context, n *node.Node, token, servicesID string, allocate bool) (chan NodeData, error) {
 	tunnels := make(chan NodeData)
 
-	err := n.Start(ctx)
-	if err != nil {
-		return nil, fmt.Errorf("creating a new node: %w", err)
-	}
 	ledger, err := n.Ledger()
 	if err != nil {
-		return nil, fmt.Errorf("creating a new node: %w", err)
+		return nil, fmt.Errorf("getting the ledger: %w", err)
 	}
 	// get new services, allocate and return to the channel
 
diff --git a/core/p2p/p2p_disabled.go b/core/p2p/p2p_disabled.go
index f0d331df..c5ba98fd 100644
--- a/core/p2p/p2p_disabled.go
+++ b/core/p2p/p2p_disabled.go
@@ -10,7 +10,7 @@ import (
 	"github.com/mudler/edgevpn/pkg/node"
 )
 
-func GenerateToken() string {
+func GenerateToken(DHTInterval, OTPInterval int) string {
 	return "not implemented"
 }
 
diff --git a/core/schema/elevenlabs.go b/core/schema/elevenlabs.go
index 8bd6be3b..119e0a58 100644
--- a/core/schema/elevenlabs.go
+++ b/core/schema/elevenlabs.go
@@ -4,3 +4,11 @@ type ElevenLabsTTSRequest struct {
 	Text    string `json:"text" yaml:"text"`
 	ModelID string `json:"model_id" yaml:"model_id"`
 }
+
+type ElevenLabsSoundGenerationRequest struct {
+	Text        string   `json:"text" yaml:"text"`
+	ModelID     string   `json:"model_id" yaml:"model_id"`
+	Duration    *float32 `json:"duration_seconds,omitempty" yaml:"duration_seconds,omitempty"`
+	Temperature *float32 `json:"prompt_influence,omitempty" yaml:"prompt_influence,omitempty"`
+	DoSample    *bool    `json:"do_sample,omitempty" yaml:"do_sample,omitempty"`
+}
diff --git a/core/schema/localai.go b/core/schema/localai.go
index 1b75e384..9070c2be 100644
--- a/core/schema/localai.go
+++ b/core/schema/localai.go
@@ -70,3 +70,7 @@ type P2PNodesResponse struct {
 	Nodes          []p2p.NodeData `json:"nodes" yaml:"nodes"`
 	FederatedNodes []p2p.NodeData `json:"federated_nodes" yaml:"federated_nodes"`
 }
+
+type SystemInformationResponse struct {
+	Backends []string `json:"backends"`
+}
diff --git a/core/services/backend_monitor.go b/core/services/backend_monitor.go
index 39588604..88fefa09 100644
--- a/core/services/backend_monitor.go
+++ b/core/services/backend_monitor.go
@@ -107,7 +107,7 @@ func (bms BackendMonitorService) CheckAndSample(modelName string) (*proto.Status
 		return nil, err
 	}
 	modelAddr := bms.modelLoader.CheckIsLoaded(backendId)
-	if modelAddr == "" {
+	if modelAddr == nil {
 		return nil, fmt.Errorf("backend %s is not currently loaded", backendId)
 	}
 
diff --git a/docs/content/docs/features/GPU-acceleration.md b/docs/content/docs/features/GPU-acceleration.md
index ae146ca6..c4160738 100644
--- a/docs/content/docs/features/GPU-acceleration.md
+++ b/docs/content/docs/features/GPU-acceleration.md
@@ -133,6 +133,10 @@ Due to the nature of ROCm it is best to run all implementations in containers as
 Ongoing verification testing of ROCm compatability with integrated backends.
 Please note the following list of verified backends and devices.
 
+LocalAI hipblas images are built against the following targets: gfx900,gfx906,gfx908,gfx940,gfx941,gfx942,gfx90a,gfx1030,gfx1031,gfx1100,gfx1101
+
+If your device is not one of these you must specify the corresponding `GPU_TARGETS` and specify `REBUILD=true`. Otherwise you don't need to specify these in the commands below.
+
 ### Verified 
 
 The devices in the following list have been tested with `hipblas` images running `ROCm 6.0.0`
diff --git a/docs/content/docs/features/distributed_inferencing.md b/docs/content/docs/features/distributed_inferencing.md
index 2de7ae3c..8c7790c6 100644
--- a/docs/content/docs/features/distributed_inferencing.md
+++ b/docs/content/docs/features/distributed_inferencing.md
@@ -68,7 +68,7 @@ And navigate the WebUI to the "Swarm" section to see the instructions to connect
 To start workers for distributing the computational load, run:
 
 ```bash
-local-ai worker llama-cpp-rpc <listening_address> <listening_port>
+local-ai worker llama-cpp-rpc --llama-cpp-args="-H <listening_address> -p <listening_port> -m <memory>" 
 ```
 
 And you can specify the address of the workers when starting LocalAI with the `LLAMACPP_GRPC_SERVERS` environment variable:
@@ -98,7 +98,7 @@ To reuse the same token later, restart the server with `--p2ptoken` or `P2P_TOKE
 2. Start the workers. Copy the `local-ai` binary to other hosts and run as many workers as needed using the token:
 
 ```bash
-TOKEN=XXX ./local-ai worker p2p-llama-cpp-rpc
+TOKEN=XXX ./local-ai worker p2p-llama-cpp-rpc --llama-cpp-args="-m <memory>" 
 # 1:06AM INF loading environment variables from file envFile=.env
 # 1:06AM INF Setting logging to info
 # {"level":"INFO","time":"2024-05-19T01:06:01.794+0200","caller":"config/config.go:288","message":"connmanager disabled\n"}
@@ -130,8 +130,10 @@ There are options that can be tweaked or parameters that can be set using enviro
 | Environment Variable | Description |
 |----------------------|-------------|
 | **LOCALAI_P2P_DISABLE_DHT** | Set to "true" to disable DHT and enable p2p layer to be local only (mDNS) |
-| **LOCALAI_P2P_DISABLE_LIMITS** | Set to "true" to disable connection limits and resources management |
+| **LOCALAI_P2P_ENABLE_LIMITS** | Set to "true" to enable connection limits and resources management (useful when running with poor connectivity or want to limit resources consumption) |
 | **LOCALAI_P2P_TOKEN** | Set the token for the p2p network |
+| **LOCALAI_P2P_LOGLEVEL** | Set the loglevel for the LocalAI p2p stack (default: info) |
+| **LOCALAI_LIBP2P_LOGLEVEL** | Set the loglevel for the underlying libp2p stack (default: fatal) |
 
 ## Architecture
 
diff --git a/docs/content/docs/features/openai-functions.md b/docs/content/docs/features/openai-functions.md
index cb667815..5d43ece0 100644
--- a/docs/content/docs/features/openai-functions.md
+++ b/docs/content/docs/features/openai-functions.md
@@ -40,43 +40,121 @@ parameters:
 To use the functions with the OpenAI client in python:
 
 ```python
-import openai
+from openai import OpenAI
+
 # ...
 # Send the conversation and available functions to GPT
-messages = [{"role": "user", "content": "What's the weather like in Boston?"}]
-functions = [
+messages = [{"role": "user", "content": "What is the weather like in Beijing now?"}]
+tools = [
     {
-        "name": "get_current_weather",
-        "description": "Get the current weather in a given location",
-        "parameters": {
-            "type": "object",
-            "properties": {
-                "location": {
-                    "type": "string",
-                    "description": "The city and state, e.g. San Francisco, CA",
+        "type": "function",
+        "function": {
+            "name": "get_current_weather",
+            "description": "Return the temperature of the specified region specified by the user",
+            "parameters": {
+                "type": "object",
+                "properties": {
+                    "location": {
+                        "type": "string",
+                        "description": "User specified region",
+                    },
+                    "unit": {
+                        "type": "string",
+                        "enum": ["celsius", "fahrenheit"],
+                        "description": "temperature unit"
+                    },
                 },
-                "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
+                "required": ["location"],
             },
-            "required": ["location"],
         },
     }
 ]
-response = openai.ChatCompletion.create(
-    model="gpt-3.5-turbo",
-    messages=messages,
-    functions=functions,
-    function_call="auto",
+
+client = OpenAI(
+    # This is the default and can be omitted
+    api_key="test",
+    base_url="http://localhost:8080/v1/"
 )
-# ...
+
+response =client.chat.completions.create(
+    messages=messages,
+    tools=tools,
+    tool_choice ="auto",
+    model="gpt-4",
+)
+#...
 ```
 
-{{% alert note %}}
-When running the python script, be sure to:
+For example, with curl:
 
-- Set `OPENAI_API_KEY` environment variable to a random string (the OpenAI api key is NOT required!)
-- Set `OPENAI_API_BASE` to point to your LocalAI service, for example `OPENAI_API_BASE=http://localhost:8080`
+```bash
+curl http://localhost:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
+  "model": "gpt-4",
+  "messages": [{"role": "user", "content": "What is the weather like in Beijing now?"}],
+  "tools": [
+        {
+            "type": "function",
+            "function": {
+                "name": "get_current_weather",
+                "description": "Return the temperature of the specified region specified by the user",
+                "parameters": {
+                    "type": "object",
+                    "properties": {
+                        "location": {
+                            "type": "string",
+                            "description": "User specified region"
+                        },
+                        "unit": {
+                            "type": "string",
+                            "enum": ["celsius", "fahrenheit"],
+                            "description": "temperature unit"
+                        }
+                    },
+                    "required": ["location"]
+                }
+            }
+        }
+    ],
+    "tool_choice":"auto"
+}'
+```
 
-{{% /alert %}}
+Return data：
+
+```json
+{
+    "created": 1724210813,
+    "object": "chat.completion",
+    "id": "16b57014-477c-4e6b-8d25-aad028a5625e",
+    "model": "gpt-4",
+    "choices": [
+        {
+            "index": 0,
+            "finish_reason": "tool_calls",
+            "message": {
+                "role": "assistant",
+                "content": "",
+                "tool_calls": [
+                    {
+                        "index": 0,
+                        "id": "16b57014-477c-4e6b-8d25-aad028a5625e",
+                        "type": "function",
+                        "function": {
+                            "name": "get_current_weather",
+                            "arguments": "{\"location\":\"Beijing\",\"unit\":\"celsius\"}"
+                        }
+                    }
+                ]
+            }
+        }
+    ],
+    "usage": {
+        "prompt_tokens": 221,
+        "completion_tokens": 26,
+        "total_tokens": 247
+    }
+}
+```
 
 ## Advanced
 
diff --git a/docs/content/docs/getting-started/quickstart.md b/docs/content/docs/getting-started/quickstart.md
index 5defa7fc..9ccc0faa 100644
--- a/docs/content/docs/getting-started/quickstart.md
+++ b/docs/content/docs/getting-started/quickstart.md
@@ -12,7 +12,7 @@ icon = "rocket_launch"
 
 **Security considerations**
 
-If you are exposing LocalAI remotely, make sure you protect the API endpoints adeguately with a mechanism which allows to protect from the incoming traffic or alternatively, run LocalAI with `API_KEY` to gate the access with an API key. The API key guarantees a total access to the features (there is no role separation), and it is to be considered as likely as an admin role.
+If you are exposing LocalAI remotely, make sure you protect the API endpoints adequately with a mechanism which allows to protect from the incoming traffic or alternatively, run LocalAI with `API_KEY` to gate the access with an API key. The API key guarantees a total access to the features (there is no role separation), and it is to be considered as likely as an admin role.
 
 To access the WebUI with an API_KEY, browser extensions such as [Requestly](https://requestly.com/) can be used (see also https://github.com/mudler/LocalAI/issues/2227#issuecomment-2093333752). See also [API flags]({{% relref "docs/advanced/advanced-usage#api-flags" %}}) for the flags / options available when starting LocalAI.
 
diff --git a/docs/content/docs/overview.md b/docs/content/docs/overview.md
index f90b1ded..5bcb6178 100644
--- a/docs/content/docs/overview.md
+++ b/docs/content/docs/overview.md
@@ -51,9 +51,7 @@ icon = "info"
 
 > 💡 Get help - [❓FAQ](https://localai.io/faq/) [💭Discussions](https://github.com/go-skynet/LocalAI/discussions) [💭Discord](https://discord.gg/uJAeKSAGDy)
 >
-> [💻 Quickstart](https://localai.io/basics/getting_started/) [📣 News](https://localai.io/basics/news/) [ 🛫 Examples ](https://github.com/go-skynet/LocalAI/tree/master/examples/) [ 🖼️ Models ](https://localai.io/models/) [ 🚀 Roadmap ](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap)
-
-
+> [💻 Quickstart](https://localai.io/basics/getting_started/) [🖼️ Models](https://models.localai.io/) [🚀 Roadmap](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap) [🥽 Demo](https://demo.localai.io) [🌍 Explorer](https://explorer.localai.io) [🛫 Examples](https://github.com/go-skynet/LocalAI/tree/master/examples/) 
 
 
 **LocalAI** is the free, Open Source OpenAI alternative. LocalAI act as a drop-in replacement REST API that's compatible with OpenAI API specifications for local inferencing. It allows you to run LLMs, generate images, audio (and not only) locally or on-prem with consumer grade hardware, supporting multiple model families and architectures. Does not require GPU. It is created and maintained by [Ettore Di Giacinto](https://github.com/mudler).
diff --git a/docs/data/version.json b/docs/data/version.json
index d07ef798..dc128c66 100644
--- a/docs/data/version.json
+++ b/docs/data/version.json
@@ -1,3 +1,3 @@
 {
-  "version": "v2.19.4"
+  "version": "v2.20.1"
 }
diff --git a/docs/themes/hugo-theme-relearn b/docs/themes/hugo-theme-relearn
index 82a5e987..d5a0ee04 160000
--- a/docs/themes/hugo-theme-relearn
+++ b/docs/themes/hugo-theme-relearn
@@ -1 +1 @@
-Subproject commit 82a5e9876c67f4c86b9e37e825e27c951ce18d54
+Subproject commit d5a0ee04ad986394d6d2f1e1a57f2334d24bf317
diff --git a/embedded/webui_static.yaml b/embedded/webui_static.yaml
index 6120ccb7..fab448cb 100644
--- a/embedded/webui_static.yaml
+++ b/embedded/webui_static.yaml
@@ -56,7 +56,4 @@
   sha: "8a9a74f4455f392ec3e7499cfda6097b536bb4b7f1e529a079c3d953c08b54ca"
 - filename: "KFOlCnqEu92Fr1MmYUtfBBc9.ttf"
   url: "https://fonts.gstatic.com/s/roboto/v30/KFOlCnqEu92Fr1MmYUtfBBc9.ttf"
-  sha: "361a50f8a6c816ba4306c5290b7e487a726e1b4dcc3d8d7e4acf1fc2dae9f551"
-- filename: "anime.js"
-  url: "https://raw.githubusercontent.com/juliangarnier/anime/master/lib/anime.min.js"
-  sha: "bceef94f964481f7680d95e7fbbe5a8c20d3945a926a754874898a578db7c7ab"
\ No newline at end of file
+  sha: "361a50f8a6c816ba4306c5290b7e487a726e1b4dcc3d8d7e4acf1fc2dae9f551"
\ No newline at end of file
diff --git a/examples/bruno/LocalAI Test Requests/Sound Generation/musicgen.bru b/examples/bruno/LocalAI Test Requests/Sound Generation/musicgen.bru
new file mode 100644
index 00000000..471756f5
--- /dev/null
+++ b/examples/bruno/LocalAI Test Requests/Sound Generation/musicgen.bru	
@@ -0,0 +1,23 @@
+meta {
+  name: musicgen
+  type: http
+  seq: 1
+}
+
+post {
+  url: {{PROTOCOL}}{{HOST}}:{{PORT}}/v1/sound-generation
+  body: json
+  auth: none
+}
+
+headers {
+  Content-Type: application/json
+}
+
+body:json {
+  {
+      "model_id": "facebook/musicgen-small",
+      "text": "Exciting 80s Newscast Interstitial",
+      "duration_seconds": 8
+  }
+}
diff --git a/examples/chainlit/requirements.txt b/examples/chainlit/requirements.txt
index 9e8b3b31..1fe9356a 100644
--- a/examples/chainlit/requirements.txt
+++ b/examples/chainlit/requirements.txt
@@ -1,6 +1,6 @@
-llama_index==0.10.65
+llama_index==0.11.9
 requests==2.32.3
-weaviate_client==4.6.7
+weaviate_client==4.8.1
 transformers
 torch
 chainlit
diff --git a/examples/functions/requirements.txt b/examples/functions/requirements.txt
index 78ffdab7..670090d3 100644
--- a/examples/functions/requirements.txt
+++ b/examples/functions/requirements.txt
@@ -1,2 +1,2 @@
-langchain==0.2.14
-openai==1.41.1
+langchain==0.3.0
+openai==1.45.1
diff --git a/examples/langchain-chroma/requirements.txt b/examples/langchain-chroma/requirements.txt
index 171ab009..4884d4aa 100644
--- a/examples/langchain-chroma/requirements.txt
+++ b/examples/langchain-chroma/requirements.txt
@@ -1,4 +1,4 @@
-langchain==0.2.14
-openai==1.40.5
+langchain==0.3.0
+openai==1.45.1
 chromadb==0.5.5
-llama-index==0.10.67.post1
\ No newline at end of file
+llama-index==0.11.7
\ No newline at end of file
diff --git a/examples/langchain/langchainpy-localai-example/requirements.txt b/examples/langchain/langchainpy-localai-example/requirements.txt
index 59330758..98325db3 100644
--- a/examples/langchain/langchainpy-localai-example/requirements.txt
+++ b/examples/langchain/langchainpy-localai-example/requirements.txt
@@ -2,23 +2,23 @@ aiohttp==3.10.3
 aiosignal==1.3.1
 async-timeout==4.0.3
 attrs==24.2.0
-certifi==2024.7.4
+certifi==2024.8.30
 charset-normalizer==3.3.2
 colorama==0.4.6
 dataclasses-json==0.6.7
 debugpy==1.8.2
 frozenlist==1.4.1
-greenlet==3.0.3
-idna==3.7
-langchain==0.2.14
-langchain-community==0.2.12
-marshmallow==3.21.3
+greenlet==3.1.0
+idna==3.8
+langchain==0.3.0
+langchain-community==0.2.16
+marshmallow==3.22.0
 marshmallow-enum==1.5.1
 multidict==6.0.5
 mypy-extensions==1.0.0
 numexpr==2.10.1
-numpy==2.1.0
-openai==1.41.1
+numpy==2.1.1
+openai==1.45.1
 openapi-schema-pydantic==1.2.4
 packaging>=23.2
 pydantic==2.8.2
@@ -30,4 +30,4 @@ tqdm==4.66.5
 typing-inspect==0.9.0
 typing_extensions==4.12.2
 urllib3==2.2.2
-yarl==1.9.4
+yarl==1.11.0
diff --git a/examples/streamlit-bot/requirements.txt b/examples/streamlit-bot/requirements.txt
index 17e1bee0..fa8c4118 100644
--- a/examples/streamlit-bot/requirements.txt
+++ b/examples/streamlit-bot/requirements.txt
@@ -1,2 +1,2 @@
-streamlit==1.37.1
+streamlit==1.38.0
 requests
\ No newline at end of file
diff --git a/gallery/hermes-vllm.yaml b/gallery/hermes-vllm.yaml
new file mode 100644
index 00000000..e8ed96b7
--- /dev/null
+++ b/gallery/hermes-vllm.yaml
@@ -0,0 +1,93 @@
+---
+name: "hermes-vllm"
+
+config_file: |
+    backend: vllm
+    parameters:
+      max_tokens: 8192
+    context_size: 8192
+    stopwords:
+    - "<|im_end|>"
+    - "<dummy32000>"
+    - "<|eot_id|>"
+    - "<|end_of_text|>"
+    function:
+      disable_no_action: true
+      grammar:
+        # Uncomment the line below to enable grammar matching for JSON results if the model is breaking
+        # the output. This will make the model more accurate and won't break the JSON output.
+        # This however, will make parallel_calls not functional (it is a known bug)
+        # mixed_mode: true
+        disable: true
+        parallel_calls: true
+        expect_strings_after_json: true
+      json_regex_match:
+      - "(?s)<tool_call>(.*?)</tool_call>"
+      - "(?s)<tool_call>(.*)"
+      capture_llm_results:
+        - (?s)<scratchpad>(.*?)</scratchpad>
+      replace_llm_results:
+        - key: (?s)<scratchpad>(.*?)</scratchpad>
+          value: ""
+
+    template:
+      use_tokenizer_template: true
+      chat: |
+        {{.Input -}}
+        <|im_start|>assistant
+      chat_message: |
+        <|im_start|>{{if eq .RoleName "assistant"}}assistant{{else if eq .RoleName "system"}}system{{else if eq .RoleName "tool"}}tool{{else if eq .RoleName "user"}}user{{end}}
+        {{- if .FunctionCall }}
+        <tool_call>
+        {{- else if eq .RoleName "tool" }}
+        <tool_response>
+        {{- end }}
+        {{- if .Content}}
+        {{.Content }}
+        {{- end }}
+        {{- if .FunctionCall}}
+        {{toJson .FunctionCall}}
+        {{- end }}
+        {{- if .FunctionCall }}
+        </tool_call>
+        {{- else if eq .RoleName "tool" }}
+        </tool_response>
+        {{- end }}<|im_end|>
+      completion: |
+        {{.Input}}
+      function: |
+        <|im_start|>system
+        You are a function calling AI model.
+        Here are the available tools:
+        <tools>
+        {{range .Functions}}
+        {'type': 'function', 'function': {'name': '{{.Name}}', 'description': '{{.Description}}', 'parameters': {{toJson .Parameters}} }}
+        {{end}}
+        </tools>
+        You should call the tools provided to you sequentially
+        Please use <scratchpad> XML tags to record your reasoning and planning before you call the functions as follows:
+        <scratchpad>
+        {step-by-step reasoning and plan in bullet points}
+        </scratchpad>
+        For each function call return a json object with function name and arguments within <tool_call> XML tags as follows:
+        <tool_call>
+        {"arguments": <args-dict>, "name": <function-name>}
+        </tool_call><|im_end|>
+        {{.Input -}}
+        <|im_start|>assistant
+# Uncomment to specify a quantization method (optional)
+# quantization: "awq"
+# Uncomment to limit the GPU memory utilization (vLLM default is 0.9 for 90%)
+# gpu_memory_utilization: 0.5
+# Uncomment to trust remote code from huggingface
+# trust_remote_code: true
+# Uncomment to enable eager execution
+# enforce_eager: true
+# Uncomment to specify the size of the CPU swap space per GPU (in GiB)
+# swap_space: 2
+# Uncomment to specify the maximum length of a sequence (including prompt and output)
+# max_model_len: 32768
+# Uncomment and specify the number of Tensor divisions.
+# Allows you to partition and run large models. Performance gains are limited.
+# https://github.com/vllm-project/vllm/issues/1435
+# tensor_parallel_size: 2
diff --git a/gallery/index.yaml b/gallery/index.yaml
index 2a10723b..229697bb 100644
--- a/gallery/index.yaml
+++ b/gallery/index.yaml
@@ -27,8 +27,8 @@
     - filename: SmolLM-1.7B-Instruct.Q4_K_M.gguf
       sha256: 2b07eb2293ed3fc544a9858beda5bfb03dcabda6aa6582d3c85768c95f498d28
       uri: huggingface://MaziyarPanahi/SmolLM-1.7B-Instruct-GGUF/SmolLM-1.7B-Instruct.Q4_K_M.gguf
-## LLama3.1
 - &llama31
+  ## LLama3.1
   url: "github:mudler/LocalAI/gallery/llama3.1-instruct.yaml@master"
   icon: https://cdn-uploads.huggingface.co/production/uploads/642cc1c253e76b4c2286c58e/aJJxKus1wP5N-euvHEUq7.png
   name: "meta-llama-3.1-8b-instruct"
@@ -85,6 +85,24 @@
     - filename: Meta-Llama-3.1-8B-Instruct.Q4_K_M.gguf
       sha256: c2f17f44af962660d1ad4cb1af91a731f219f3b326c2b14441f9df1f347f2815
       uri: huggingface://MaziyarPanahi/Meta-Llama-3.1-8B-Instruct-GGUF/Meta-Llama-3.1-8B-Instruct.Q4_K_M.gguf
+- !!merge <<: *llama31
+  name: "meta-llama-3.1-8b-instruct:Q8_grammar-functioncall"
+  url: "github:mudler/LocalAI/gallery/llama3.1-instruct-grammar.yaml@master"
+  urls:
+    - https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct
+    - https://huggingface.co/MaziyarPanahi/Meta-Llama-3.1-8B-Instruct-GGUF
+  description: |
+    This is the standard Llama 3.1 8B Instruct model with grammar and function call enabled.
+
+    When grammars are enabled in LocalAI, the LLM is forced to output valid tools constrained by BNF grammars. This can be useful for ensuring that the model outputs are valid and can be used in a production environment.
+    For more information on how to use grammars in LocalAI, see https://localai.io/features/openai-functions/#advanced and https://localai.io/features/constrained_grammars/.
+  overrides:
+    parameters:
+      model: Meta-Llama-3.1-8B-Instruct.Q8_0.gguf
+  files:
+    - filename: Meta-Llama-3.1-8B-Instruct.Q8_0.gguf
+      sha256: f8d608c983b83a1bf28229bc9beb4294c91f5d4cbfe2c1829566b4d7c4693eeb
+      uri: huggingface://MaziyarPanahi/Meta-Llama-3.1-8B-Instruct-GGUF/Meta-Llama-3.1-8B-Instruct.Q8_0.gguf
 - !!merge <<: *llama31
   name: "meta-llama-3.1-8b-claude-imat"
   urls:
@@ -389,6 +407,38 @@
     - filename: llama-3.1-storm-8b-q4_k_m.gguf
       sha256: d714e960211ee0fe6113d3131a6573e438f37debd07e1067d2571298624414a0
       uri: huggingface://mudler/Llama-3.1-Storm-8B-Q4_K_M-GGUF/llama-3.1-storm-8b-q4_k_m.gguf
+- !!merge <<: *llama31
+  name: "hubble-4b-v1"
+  url: "github:mudler/LocalAI/gallery/chatml.yaml@master"
+  icon: https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/R8_o3CCpTgKv5Wnnry7E_.png
+  urls:
+    - https://huggingface.co/TheDrummer/Hubble-4B-v1-GGUF
+  description: |
+    Equipped with his five senses, man explores the universe around him and calls the adventure 'Science'.
+    This is a finetune of Nvidia's Llama 3.1 4B Minitron - a shrunk down model of Llama 3.1 8B 128K.
+  overrides:
+    parameters:
+      model: Hubble-4B-v1-Q4_K_M.gguf
+  files:
+    - filename: Hubble-4B-v1-Q4_K_M.gguf
+      uri: huggingface://TheDrummer/Hubble-4B-v1-GGUF/Hubble-4B-v1-Q4_K_M.gguf
+      sha256: 0721294d0e861c6e6162a112fc7242e0c4b260c156137f4bcbb08667f1748080
+- !!merge <<: *llama31
+  name: "reflection-llama-3.1-70b"
+  urls:
+    - https://huggingface.co/leafspark/Reflection-Llama-3.1-70B-bf16
+    - https://huggingface.co/senseable/Reflection-Llama-3.1-70B-gguf
+  description: |
+      Reflection Llama-3.1 70B is (currently) the world's top open-source LLM, trained with a new technique called Reflection-Tuning that teaches a LLM to detect mistakes in its reasoning and correct course.
+
+      The model was trained on synthetic data generated by Glaive. If you're training a model, Glaive is incredible — use them.
+  overrides:
+    parameters:
+      model: Reflection-Llama-3.1-70B-q4_k_m.gguf
+  files:
+    - filename: Reflection-Llama-3.1-70B-q4_k_m.gguf
+      sha256: 16064e07037883a750cfeae9a7be41143aa857dbac81c2e93c68e2f941dee7b2
+      uri: huggingface://senseable/Reflection-Llama-3.1-70B-gguf/Reflection-Llama-3.1-70B-q4_k_m.gguf
 ## Uncensored models
 - !!merge <<: *llama31
   name: "humanish-roleplay-llama-3.1-8b-i1"
@@ -608,6 +658,89 @@
     - filename: Mahou-1.3-llama3.1-8B.Q4_K_M.gguf
       sha256: 88bfdca2f6077d789d3e0f161d19711aa208a6d9a02cce96a2276c69413b3594
       uri: huggingface://mradermacher/Mahou-1.3-llama3.1-8B-GGUF/Mahou-1.3-llama3.1-8B.Q4_K_M.gguf
+- !!merge <<: *llama31
+  name: "azure_dusk-v0.2-iq-imatrix"
+  # chatml
+  url: "github:mudler/LocalAI/gallery/chatml.yaml@master"
+  icon: https://cdn-uploads.huggingface.co/production/uploads/65d4cf2693a0a3744a27536c/n3-g_YTk3FY-DBzxXd28E.png
+  urls:
+    - https://huggingface.co/Lewdiculous/Azure_Dusk-v0.2-GGUF-IQ-Imatrix
+  description: |
+    "Following up on Crimson_Dawn-v0.2 we have Azure_Dusk-v0.2! Training on Mistral-Nemo-Base-2407 this time I've added significantly more data, as well as trained using RSLoRA as opposed to regular LoRA. Another key change is training on ChatML as opposed to Mistral Formatting."
+    by Author.
+  overrides:
+    parameters:
+      model: Azure_Dusk-v0.2-Q4_K_M-imat.gguf
+  files:
+    - filename: Azure_Dusk-v0.2-Q4_K_M-imat.gguf
+      sha256: c03a670c00976d14c267a0322374ed488b2a5f4790eb509136ca4e75cbc10cf4
+      uri: huggingface://Lewdiculous/Azure_Dusk-v0.2-GGUF-IQ-Imatrix/Azure_Dusk-v0.2-Q4_K_M-imat.gguf
+- !!merge <<: *llama31
+  name: "l3.1-8b-niitama-v1.1-iq-imatrix"
+  icon: https://cdn-uploads.huggingface.co/production/uploads/65d4cf2693a0a3744a27536c/2Q5ky8TvP0vLS1ulMXnrn.png
+  urls:
+    - https://huggingface.co/Sao10K/L3.1-8B-Niitama-v1.1
+    - https://huggingface.co/Lewdiculous/L3.1-8B-Niitama-v1.1-GGUF-IQ-Imatrix
+  description: |
+   GGUF-IQ-Imatrix quants for Sao10K/L3.1-8B-Niitama-v1.1
+   Here's the subjectively superior L3 version: L3-8B-Niitama-v1
+   An experimental model using experimental methods.
+
+   More detail on it:
+
+   Tamamo and Niitama are made from the same data. Literally. The only thing that's changed is how theyre shuffled and formatted. Yet, I get wildly different results.
+
+   Interesting, eh? Feels kinda not as good compared to the l3 version, but it's aight.
+  overrides:
+    parameters:
+      model: L3.1-8B-Niitama-v1.1-Q4_K_M-imat.gguf
+  files:
+    - filename: L3.1-8B-Niitama-v1.1-Q4_K_M-imat.gguf
+      sha256: 524163bd0f1d43c9284b09118abcc192f3250b13dd3bb79d60c28321108b6748
+      uri: huggingface://Lewdiculous/L3.1-8B-Niitama-v1.1-GGUF-IQ-Imatrix/L3.1-8B-Niitama-v1.1-Q4_K_M-imat.gguf
+- !!merge <<: *llama31
+  name: "llama-3.1-8b-stheno-v3.4-iq-imatrix"
+  icon: https://huggingface.co/Sao10K/Llama-3.1-8B-Stheno-v3.4/resolve/main/meneno.jpg
+  urls:
+    - https://huggingface.co/Sao10K/Llama-3.1-8B-Stheno-v3.4
+    - https://huggingface.co/Lewdiculous/Llama-3.1-8B-Stheno-v3.4-GGUF-IQ-Imatrix
+  description: |
+    This model has went through a multi-stage finetuning process.
+
+    - 1st, over a multi-turn Conversational-Instruct
+    - 2nd, over a Creative Writing / Roleplay along with some Creative-based Instruct Datasets.
+    - - Dataset consists of a mixture of Human and Claude Data.
+
+    Prompting Format:
+
+    - Use the L3 Instruct Formatting - Euryale 2.1 Preset Works Well
+    - Temperature + min_p as per usual, I recommend 1.4 Temp + 0.2 min_p.
+    - Has a different vibe to previous versions. Tinker around.
+
+    Changes since previous Stheno Datasets:
+
+    - Included Multi-turn Conversation-based Instruct Datasets to boost multi-turn coherency. # This is a seperate set, not the ones made by Kalomaze and Nopm, that are used in Magnum. They're completely different data.
+    - Replaced Single-Turn Instruct with Better Prompts and Answers by Claude 3.5 Sonnet and Claude 3 Opus.
+    - Removed c2 Samples -> Underway of re-filtering and masking to use with custom prefills. TBD
+    - Included 55% more Roleplaying Examples based of [Gryphe's](https://huggingface.co/datasets/Gryphe/Sonnet3.5-Charcard-Roleplay) Charcard RP Sets. Further filtered and cleaned on.
+    - Included 40% More Creative Writing Examples.
+    - Included Datasets Targeting System Prompt Adherence.
+    - Included Datasets targeting Reasoning / Spatial Awareness.
+    - Filtered for the usual errors, slop and stuff at the end. Some may have slipped through, but I removed nearly all of it.
+
+    Personal Opinions:
+
+    - Llama3.1 was more disappointing, in the Instruct Tune? It felt overbaked, atleast. Likely due to the DPO being done after their SFT Stage.
+    - Tuning on L3.1 base did not give good results, unlike when I tested with Nemo base. unfortunate.
+    - Still though, I think I did an okay job. It does feel a bit more distinctive.
+    - It took a lot of tinkering, like a LOT to wrangle this.
+  overrides:
+    parameters:
+      model: Llama-3.1-8B-Stheno-v3.4-Q4_K_M-imat.gguf
+  files:
+    - filename: Llama-3.1-8B-Stheno-v3.4-Q4_K_M-imat.gguf
+      sha256: 830d4858aa11a654f82f69fa40dee819edf9ecf54213057648304eb84b8dd5eb
+      uri: huggingface://Lewdiculous/Llama-3.1-8B-Stheno-v3.4-GGUF-IQ-Imatrix/Llama-3.1-8B-Stheno-v3.4-Q4_K_M-imat.gguf
 - &deepseek
   ## Deepseek
   url: "github:mudler/LocalAI/gallery/deepseek.yaml@master"
@@ -1114,7 +1247,7 @@
     - https://huggingface.co/TheDrummer/Rocinante-12B-v1.1-GGUF
     - https://huggingface.co/TheDrummer/Rocinante-12B-v1.1
   description: |
-     A versatile workhorse for any adventure!
+    A versatile workhorse for any adventure!
   overrides:
     parameters:
       model: Rocinante-12B-v1.1-Q4_K_M.gguf
@@ -1145,6 +1278,23 @@
     - filename: Pantheon-RP-1.6-12b-Nemo-Q4_K_M.gguf
       sha256: cf3465c183bf4ecbccd1b6b480f687e0160475b04c87e2f1e5ebc8baa0f4c7aa
       uri: huggingface://bartowski/Pantheon-RP-1.6-12b-Nemo-GGUF/Pantheon-RP-1.6-12b-Nemo-Q4_K_M.gguf
+- !!merge <<: *mistral03
+  name: "mn-12b-lyra-v4-iq-imatrix"
+  icon: https://cdn-uploads.huggingface.co/production/uploads/65d4cf2693a0a3744a27536c/dVoru83WOpwVjMlgZ_xhA.png
+  # chatml
+  url: "github:mudler/LocalAI/gallery/chatml.yaml@master"
+  urls:
+    - https://huggingface.co/Lewdiculous/MN-12B-Lyra-v4-GGUF-IQ-Imatrix
+  description: |
+      A finetune of Mistral Nemo by Sao10K.
+      Uses the ChatML prompt format.
+  overrides:
+    parameters:
+      model: MN-12B-Lyra-v4-Q4_K_M-imat.gguf
+  files:
+    - filename: MN-12B-Lyra-v4-Q4_K_M-imat.gguf
+      sha256: 1989123481ca1936c8a2cbe278ff5d1d2b0ae63dbdc838bb36a6d7547b8087b3
+      uri: huggingface://Lewdiculous/MN-12B-Lyra-v4-GGUF-IQ-Imatrix/MN-12B-Lyra-v4-Q4_K_M-imat.gguf
 - &mudler
   ### START mudler's LocalAI specific-models
   url: "github:mudler/LocalAI/gallery/mudler.yaml@master"
@@ -1658,6 +1808,48 @@
     - filename: shieldgemma-9b.i1-Q4_K_M.gguf
       sha256: ffa7eaadcc0c7d0544fda5b0d86bba3ffa3431b673e5b2135f421cfe65bd8732
       uri: huggingface://mradermacher/shieldgemma-9b-i1-GGUF/shieldgemma-9b.i1-Q4_K_M.gguf
+- !!merge <<: *gemma
+  name: "athena-codegemma-2-2b-it"
+  urls:
+    - https://huggingface.co/EpistemeAI/Athena-codegemma-2-2b-it
+    - https://huggingface.co/mradermacher/Athena-codegemma-2-2b-it-GGUF
+  description: |
+      Supervised fine tuned (sft unsloth) for coding with EpistemeAI coding dataset.
+  overrides:
+    parameters:
+      model: Athena-codegemma-2-2b-it.Q4_K_M.gguf
+  files:
+    - filename: Athena-codegemma-2-2b-it.Q4_K_M.gguf
+      sha256: 59ce17023438b0da603dd211c7d39f78e7acac4108258ac0818a97a4ca7d64e3
+      uri: huggingface://mradermacher/Athena-codegemma-2-2b-it-GGUF/Athena-codegemma-2-2b-it.Q4_K_M.gguf
+- !!merge <<: *gemma
+  name: "datagemma-rag-27b-it"
+  urls:
+    - https://huggingface.co/google/datagemma-rag-27b-it
+    - https://huggingface.co/bartowski/datagemma-rag-27b-it-GGUF
+  description: |
+    DataGemma is a series of fine-tuned Gemma 2 models used to help LLMs access and incorporate reliable public statistical data from Data Commons into their responses. DataGemma RAG is used with Retrieval Augmented Generation, where it is trained to take a user query and generate natural language queries that can be understood by Data Commons' existing natural language interface. More information can be found in this research paper.
+  overrides:
+    parameters:
+      model: datagemma-rag-27b-it-Q4_K_M.gguf
+  files:
+    - filename: datagemma-rag-27b-it-Q4_K_M.gguf
+      sha256: 3dfcf51b05e3f0ab0979ad194de350edea71cb14444efa0a9f2ef5bfc80753f8
+      uri: huggingface://bartowski/datagemma-rag-27b-it-GGUF/datagemma-rag-27b-it-Q4_K_M.gguf
+- !!merge <<: *gemma
+  name: "datagemma-rig-27b-it"
+  urls:
+    - https://huggingface.co/google/datagemma-rig-27b-it
+    - https://huggingface.co/bartowski/datagemma-rig-27b-it-GGUF
+  description: |
+    DataGemma is a series of fine-tuned Gemma 2 models used to help LLMs access and incorporate reliable public statistical data from Data Commons into their responses. DataGemma RIG is used in the retrieval interleaved generation approach (based off of tool-use approaches), where it is trained to annotate a response with natural language queries to Data Commons’ existing natural language interface wherever there are statistics. More information can be found in this research paper.
+  overrides:
+    parameters:
+      model: datagemma-rig-27b-it-Q4_K_M.gguf
+  files:
+    - filename: datagemma-rig-27b-it-Q4_K_M.gguf
+      sha256: a6738ffbb49b6c46d220e2793df85c0538e9ac72398e32a0914ee5e55c3096ad
+      uri: huggingface://bartowski/datagemma-rig-27b-it-GGUF/datagemma-rig-27b-it-Q4_K_M.gguf
 - &llama3
   url: "github:mudler/LocalAI/gallery/llama3-instruct.yaml@master"
   icon: https://cdn-uploads.huggingface.co/production/uploads/642cc1c253e76b4c2286c58e/aJJxKus1wP5N-euvHEUq7.png
@@ -1690,7 +1882,7 @@
   files:
     - filename: Meta-Llama-3-8B-Instruct.Q4_0.gguf
       uri: huggingface://QuantFactory/Meta-Llama-3-8B-Instruct-GGUF/Meta-Llama-3-8B-Instruct.Q4_0.gguf
-      sha256: 1977ae6185ef5bc476e27db85bb3d79ca4bd87e7b03399083c297d9c612d334c
+      sha256: 2b4675c2208f09ad8762d8cf1b6a4a26bf65e6f0641aba324ec65143c0b4ad9f
 - !!merge <<: *llama3
   name: "llama3-8b-instruct:Q6_K"
   overrides:
@@ -1699,7 +1891,7 @@
   files:
     - filename: Meta-Llama-3-8B-Instruct.Q6_K.gguf
       uri: huggingface://QuantFactory/Meta-Llama-3-8B-Instruct-GGUF/Meta-Llama-3-8B-Instruct.Q6_K.gguf
-      sha256: d1cdc49a716674c8e2506039bef85b905376cdaafc33e449b5aa2cf88f9532ad
+      sha256: bd7efd73f9fb67e4b9ecc43f861f37c7e594e78a8a5ff9c29da021692bd243ef
 - !!merge <<: *llama3
   name: "llama-3-8b-instruct-abliterated"
   urls:
@@ -3142,6 +3334,121 @@
     - filename: Master-Yi-9B_Q4_K_M.gguf
       sha256: 57e2afcf9f24d7138a3b8e2b547336d7edc13621a5e8090bc196d7de360b2b45
       uri: huggingface://qnguyen3/Master-Yi-9B-GGUF/Master-Yi-9B_Q4_K_M.gguf
+- !!merge <<: *yi-chat
+  name: "magnum-v3-34b"
+  icon: https://cdn-uploads.huggingface.co/production/uploads/658a46cbfb9c2bdfae75b3a6/9yEmnTDG9bcC_bxwuDU6G.png
+  urls:
+    - https://huggingface.co/anthracite-org/magnum-v3-34b
+    - https://huggingface.co/bartowski/magnum-v3-34b-GGUF
+  description: |
+    This is the 9th in a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet and Opus.
+
+    This model is fine-tuned on top of Yi-1.5-34 B-32 K.
+  overrides:
+    parameters:
+      model: magnum-v3-34b-Q4_K_M.gguf
+  files:
+    - filename: magnum-v3-34b-Q4_K_M.gguf
+      sha256: f902956c0731581f1ff189e547e6e5aad86b77af5f4dc7e4fc26bcda5c1f7cc3
+      uri: huggingface://bartowski/magnum-v3-34b-GGUF/magnum-v3-34b-Q4_K_M.gguf
+- !!merge <<: *yi-chat
+  name: "yi-coder-9b-chat"
+  urls:
+    - https://huggingface.co/01-ai/Yi-Coder-9B-Chat
+    - https://huggingface.co/bartowski/Yi-Coder-9B-Chat-GGUF
+    - https://01-ai.github.io/
+    - https://github.com/01-ai/Yi-Coder
+  description: |
+    Yi-Coder is a series of open-source code language models that delivers state-of-the-art coding performance with fewer than 10 billion parameters.
+    Key features:
+
+        Excelling in long-context understanding with a maximum context length of 128K tokens.
+        Supporting 52 major programming languages:
+
+      'java', 'markdown', 'python', 'php', 'javascript', 'c++', 'c#', 'c', 'typescript', 'html', 'go', 'java_server_pages', 'dart', 'objective-c', 'kotlin', 'tex', 'swift', 'ruby', 'sql', 'rust', 'css', 'yaml', 'matlab', 'lua', 'json', 'shell', 'visual_basic', 'scala', 'rmarkdown', 'pascal', 'fortran', 'haskell', 'assembly', 'perl', 'julia', 'cmake', 'groovy', 'ocaml', 'powershell', 'elixir', 'clojure', 'makefile', 'coffeescript', 'erlang', 'lisp', 'toml', 'batchfile', 'cobol', 'dockerfile', 'r', 'prolog', 'verilog'
+
+    For model details and benchmarks, see Yi-Coder blog and Yi-Coder README.
+  overrides:
+    parameters:
+      model: Yi-Coder-9B-Chat-Q4_K_M.gguf
+  files:
+    - filename: Yi-Coder-9B-Chat-Q4_K_M.gguf
+      sha256: 251cc196e3813d149694f362bb0f8f154f3320abe44724eebe58c23dc54f201d
+      uri: huggingface://bartowski/Yi-Coder-9B-Chat-GGUF/Yi-Coder-9B-Chat-Q4_K_M.gguf
+- !!merge <<: *yi-chat
+  name: "yi-coder-1.5b-chat"
+  urls:
+    - https://huggingface.co/01-ai/Yi-Coder-1.5B-Chat
+    - https://huggingface.co/MaziyarPanahi/Yi-Coder-1.5B-Chat-GGUF
+    - https://01-ai.github.io/
+    - https://github.com/01-ai/Yi-Coder
+  description: |
+    Yi-Coder is a series of open-source code language models that delivers state-of-the-art coding performance with fewer than 10 billion parameters.
+    Key features:
+
+        Excelling in long-context understanding with a maximum context length of 128K tokens.
+        Supporting 52 major programming languages:
+
+      'java', 'markdown', 'python', 'php', 'javascript', 'c++', 'c#', 'c', 'typescript', 'html', 'go', 'java_server_pages', 'dart', 'objective-c', 'kotlin', 'tex', 'swift', 'ruby', 'sql', 'rust', 'css', 'yaml', 'matlab', 'lua', 'json', 'shell', 'visual_basic', 'scala', 'rmarkdown', 'pascal', 'fortran', 'haskell', 'assembly', 'perl', 'julia', 'cmake', 'groovy', 'ocaml', 'powershell', 'elixir', 'clojure', 'makefile', 'coffeescript', 'erlang', 'lisp', 'toml', 'batchfile', 'cobol', 'dockerfile', 'r', 'prolog', 'verilog'
+
+    For model details and benchmarks, see Yi-Coder blog and Yi-Coder README.
+  overrides:
+    parameters:
+      model: Yi-Coder-1.5B-Chat.Q4_K_M.gguf
+  files:
+    - filename: Yi-Coder-1.5B-Chat.Q4_K_M.gguf
+      sha256: e2e8fa659cd75c828d7783b5c2fb60d220e08836065901fad8edb48e537c1cec
+      uri: huggingface://MaziyarPanahi/Yi-Coder-1.5B-Chat-GGUF/Yi-Coder-1.5B-Chat.Q4_K_M.gguf
+- !!merge <<: *yi-chat
+  url: "github:mudler/LocalAI/gallery/codellama.yaml@master"
+  name: "yi-coder-1.5b"
+  urls:
+    - https://huggingface.co/01-ai/Yi-Coder-1.5B
+    - https://huggingface.co/QuantFactory/Yi-Coder-1.5B-GGUF
+    - https://01-ai.github.io/
+    - https://github.com/01-ai/Yi-Coder
+  description: |
+    Yi-Coder is a series of open-source code language models that delivers state-of-the-art coding performance with fewer than 10 billion parameters.
+    Key features:
+
+        Excelling in long-context understanding with a maximum context length of 128K tokens.
+        Supporting 52 major programming languages:
+
+      'java', 'markdown', 'python', 'php', 'javascript', 'c++', 'c#', 'c', 'typescript', 'html', 'go', 'java_server_pages', 'dart', 'objective-c', 'kotlin', 'tex', 'swift', 'ruby', 'sql', 'rust', 'css', 'yaml', 'matlab', 'lua', 'json', 'shell', 'visual_basic', 'scala', 'rmarkdown', 'pascal', 'fortran', 'haskell', 'assembly', 'perl', 'julia', 'cmake', 'groovy', 'ocaml', 'powershell', 'elixir', 'clojure', 'makefile', 'coffeescript', 'erlang', 'lisp', 'toml', 'batchfile', 'cobol', 'dockerfile', 'r', 'prolog', 'verilog'
+
+    For model details and benchmarks, see Yi-Coder blog and Yi-Coder README.
+  overrides:
+    parameters:
+      model: Yi-Coder-1.5B.Q4_K_M.gguf
+  files:
+    - filename: Yi-Coder-1.5B.Q4_K_M.gguf
+      sha256: 86a280dd36c9b2342b7023532f9c2c287e251f5cd10bc81ca262db8c1668f272
+      uri: huggingface://QuantFactory/Yi-Coder-1.5B-GGUF/Yi-Coder-1.5B.Q4_K_M.gguf
+- !!merge <<: *yi-chat
+  url: "github:mudler/LocalAI/gallery/codellama.yaml@master"
+  name: "yi-coder-9b"
+  urls:
+    - https://huggingface.co/01-ai/Yi-Coder-9B
+    - https://huggingface.co/QuantFactory/Yi-Coder-9B-GGUF
+    - https://01-ai.github.io/
+    - https://github.com/01-ai/Yi-Coder
+  description: |
+    Yi-Coder is a series of open-source code language models that delivers state-of-the-art coding performance with fewer than 10 billion parameters.
+    Key features:
+
+        Excelling in long-context understanding with a maximum context length of 128K tokens.
+        Supporting 52 major programming languages:
+
+      'java', 'markdown', 'python', 'php', 'javascript', 'c++', 'c#', 'c', 'typescript', 'html', 'go', 'java_server_pages', 'dart', 'objective-c', 'kotlin', 'tex', 'swift', 'ruby', 'sql', 'rust', 'css', 'yaml', 'matlab', 'lua', 'json', 'shell', 'visual_basic', 'scala', 'rmarkdown', 'pascal', 'fortran', 'haskell', 'assembly', 'perl', 'julia', 'cmake', 'groovy', 'ocaml', 'powershell', 'elixir', 'clojure', 'makefile', 'coffeescript', 'erlang', 'lisp', 'toml', 'batchfile', 'cobol', 'dockerfile', 'r', 'prolog', 'verilog'
+
+    For model details and benchmarks, see Yi-Coder blog and Yi-Coder README.
+  overrides:
+    parameters:
+      model: Yi-Coder-9B.Q4_K_M.gguf
+  files:
+    - filename: Yi-Coder-9B.Q4_K_M.gguf
+      sha256: cff3db8a69c43654e3c2d2984e86ad2791d1d446ec56b24a636ba1ce78363308
+      uri: huggingface://QuantFactory/Yi-Coder-9B-GGUF/Yi-Coder-9B.Q4_K_M.gguf
 - &vicuna-chat
   ## LLama2 and derivatives
   ### Start Fimbulvetr
@@ -4555,7 +4862,7 @@
     - filename: phillama-3.8b-v0.1.Q4_K_M.gguf
       sha256: da537d352b7aae54bbad0d2cff3e3a1b0e1dc1e1d25bec3aae1d05cf4faee7a2
       uri: huggingface://RichardErkhov/raincandy-u_-_phillama-3.8b-v0.1-gguf/phillama-3.8b-v0.1.Q4_K_M.gguf
-- !!merge <<: *llama3
+- !!merge <<: *phi-3
   name: "calme-2.3-phi3-4b"
   icon: https://huggingface.co/MaziyarPanahi/calme-2.1-phi3-4b/resolve/main/phi-3-instruct.webp
   urls:
@@ -4572,6 +4879,35 @@
     - filename: Phi-3-mini-4k-instruct-v0.3.Q4_K_M.gguf
       sha256: 3a23e1052369c080afb925882bd814cbea5ec859894655a7434c3d49e43a6127
       uri: huggingface://MaziyarPanahi/calme-2.3-phi3-4b-GGUF/Phi-3-mini-4k-instruct-v0.3.Q4_K_M.gguf
+- !!merge <<: *phi-3
+  name: "phi-3.5-mini-instruct"
+  urls:
+    - https://huggingface.co/microsoft/Phi-3.5-mini-instruct
+    - https://huggingface.co/MaziyarPanahi/Phi-3.5-mini-instruct-GGUF
+  description: |
+    Phi-3.5-mini is a lightweight, state-of-the-art open model built upon datasets used for Phi-3 - synthetic data and filtered publicly available websites - with a focus on very high-quality, reasoning dense data. The model belongs to the Phi-3 model family and supports 128K token context length. The model underwent a rigorous enhancement process, incorporating both supervised fine-tuning, proximal policy optimization, and direct preference optimization to ensure precise instruction adherence and robust safety measures.
+  overrides:
+    parameters:
+      model: Phi-3.5-mini-instruct.Q4_K_M.gguf
+  files:
+    - filename: Phi-3.5-mini-instruct.Q4_K_M.gguf
+      sha256: 3f68916e850b107d8641d18bcd5548f0d66beef9e0a9077fe84ef28943eb7e88
+      uri: huggingface://MaziyarPanahi/Phi-3.5-mini-instruct-GGUF/Phi-3.5-mini-instruct.Q4_K_M.gguf
+- !!merge <<: *phi-3
+  name: "calme-2.1-phi3.5-4b-i1"
+  icon: https://huggingface.co/MaziyarPanahi/calme-2.1-phi3.5-4b/resolve/main/calme-2.webp
+  urls:
+    - https://huggingface.co/MaziyarPanahi/calme-2.1-phi3.5-4b
+    - https://huggingface.co/mradermacher/calme-2.1-phi3.5-4b-i1-GGUF
+  description: |
+    This model is a fine-tuned version of the microsoft/Phi-3.5-mini-instruct, pushing the boundaries of natural language understanding and generation even further. My goal was to create a versatile and robust model that excels across a wide range of benchmarks and real-world applications.
+  overrides:
+    parameters:
+      model: calme-2.1-phi3.5-4b.i1-Q4_K_M.gguf
+  files:
+    - filename: calme-2.1-phi3.5-4b.i1-Q4_K_M.gguf
+      sha256: 989eccacd52b6d9ebf2c06c35c363da19aadb125659a10df299b7130bc293e77
+      uri: huggingface://mradermacher/calme-2.1-phi3.5-4b-i1-GGUF/calme-2.1-phi3.5-4b.i1-Q4_K_M.gguf
 - &hermes-2-pro-mistral
   ### START Hermes
   url: "github:mudler/LocalAI/gallery/hermes-2-pro-mistral.yaml@master"
@@ -4737,6 +5073,21 @@
     - filename: Hermes-3-Llama-3.1-8B.Q4_K_M.gguf
       sha256: d4403ce5a6e930f4c2509456388c20d633a15ff08dd52ef3b142ff1810ec3553
       uri: huggingface://NousResearch/Hermes-3-Llama-3.1-8B-GGUF/Hermes-3-Llama-3.1-8B.Q4_K_M.gguf
+- !!merge <<: *hermes-2-pro-mistral
+  name: "hermes-3-llama-3.1-8b:Q8"
+  icon: https://cdn-uploads.huggingface.co/production/uploads/6317aade83d8d2fd903192d9/bMcZ3sNNQK8SRZpHXBmwM.jpeg
+  urls:
+    - https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B
+    - https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B-GGUF
+  description: |
+    Hermes 3 is a generalist language model developed by Nous Research. It is an advanced agentic model with improved roleplaying, reasoning, multi-turn conversation, long context coherence, and generalist assistant capabilities. The model is built on top of the Llama-3 architecture and has been fine-tuned to achieve superior performance in various tasks. It is designed to be a powerful and reliable tool for solving complex problems and assisting users in achieving their goals. Hermes 3 can be used for a wide range of applications, including research, education, and personal assistant tasks. It is available on the Hugging Face model hub for easy access and integration into existing workflows.
+  overrides:
+    parameters:
+      model: Hermes-3-Llama-3.1-8B.Q8_0.gguf
+  files:
+    - filename: Hermes-3-Llama-3.1-8B.Q8_0.gguf
+      sha256: c77c263f78b2f56fbaddd3ef2af750fda6ebb4344a546aaa0bfdd546b1ca8d84
+      uri: huggingface://NousResearch/Hermes-3-Llama-3.1-8B-GGUF/Hermes-3-Llama-3.1-8B.Q8_0.gguf
 - !!merge <<: *hermes-2-pro-mistral
   name: "hermes-3-llama-3.1-70b"
   icon: https://cdn-uploads.huggingface.co/production/uploads/6317aade83d8d2fd903192d9/vG6j5WxHX09yj32vgjJlI.jpeg
@@ -4752,6 +5103,53 @@
     - filename: Hermes-3-Llama-3.1-70B.Q4_K_M.gguf
       sha256: 955c2f42caade4278f3c9dbffa32bb74572652b20e49e5340e782de3585bbe3f
       uri: huggingface://NousResearch/Hermes-3-Llama-3.1-70B-GGUF/Hermes-3-Llama-3.1-70B.Q4_K_M.gguf
+- !!merge <<: *hermes-2-pro-mistral
+  name: "hermes-3-llama-3.1-70b:Q5_K_M"
+  icon: https://cdn-uploads.huggingface.co/production/uploads/6317aade83d8d2fd903192d9/vG6j5WxHX09yj32vgjJlI.jpeg
+  urls:
+    - https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-70B
+    - https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-70B-GGUF
+  description: |
+    Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the board. It is designed to focus on aligning LLMs to the user, with powerful steering capabilities and control given to the end user. The model uses ChatML as the prompt format, opening up a much more structured system for engaging the LLM in multi-turn chat dialogue. It also supports function calling and structured output capabilities, generalist assistant capabilities, and improved code generation skills.
+  overrides:
+    parameters:
+      model: Hermes-3-Llama-3.1-70B.Q5_K_M.gguf
+  files:
+    - filename: Hermes-3-Llama-3.1-70B.Q5_K_M.gguf
+      sha256: 10ae3e0441b14c4a6476436f3c14e8bcacc7928aa3e8ce978d053287289a7ebb
+      uri: huggingface://NousResearch/Hermes-3-Llama-3.1-70B-GGUF/Hermes-3-Llama-3.1-70B.Q5_K_M.gguf
+- &hermes-vllm
+  url: "github:mudler/LocalAI/gallery/hermes-vllm.yaml@master"
+  name: "hermes-3-llama-3.1-8b:vllm"
+  icon: https://cdn-uploads.huggingface.co/production/uploads/6317aade83d8d2fd903192d9/vG6j5WxHX09yj32vgjJlI.jpeg
+  tags:
+    - llm
+    - vllm
+    - gpu
+    - function-calling
+  license: llama-3
+  urls:
+    - https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B
+  description: |
+    Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the board. It is designed to focus on aligning LLMs to the user, with powerful steering capabilities and control given to the end user. The model uses ChatML as the prompt format, opening up a much more structured system for engaging the LLM in multi-turn chat dialogue. It also supports function calling and structured output capabilities, generalist assistant capabilities, and improved code generation skills.
+  overrides:
+    parameters:
+      model: NousResearch/Hermes-3-Llama-3.1-8B
+- !!merge <<: *hermes-vllm
+  name: "hermes-3-llama-3.1-70b:vllm"
+  urls:
+    - https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-70B
+  overrides:
+    parameters:
+      model: NousResearch/Hermes-3-Llama-3.1-70B
+- !!merge <<: *hermes-vllm
+  name: "hermes-3-llama-3.1-405b:vllm"
+  icon: https://cdn-uploads.huggingface.co/production/uploads/6317aade83d8d2fd903192d9/-kj_KflXsdpcZoTQsvx7W.jpeg
+  urls:
+    - https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-405B
+  overrides:
+    parameters:
+      model: NousResearch/Hermes-3-Llama-3.1-405B
 - !!merge <<: *hermes-2-pro-mistral
   name: "biomistral-7b"
   description: |
@@ -5745,6 +6143,16 @@
     - filename: voice-it-riccardo_fasol-x-low.tar.gz
       uri: https://github.com/rhasspy/piper/releases/download/v0.0.2/voice-it-riccardo_fasol-x-low.tar.gz
       sha256: 394b27b8780f5167e73a62ac103839cc438abc7edb544192f965e5b8f5f4acdb
+- !!merge <<: *piper
+  url: github:mudler/LocalAI/gallery/piper.yaml@master
+  name: voice-it-paola-medium
+  overrides:
+    parameters:
+      model: it-paola-medium.onnx
+  files:
+    - filename: voice-it-paola-medium.tar.gz
+      uri: https://github.com/fakezeta/piper-paola-voice/releases/download/v1.0.0/voice-it-paola-medium.tar.gz
+      sha256: 61d3bac0ff6d347daea5464c4b3ae156a450b603a916cc9ed7deecdeba17153a
 - !!merge <<: *piper
   url: github:mudler/LocalAI/gallery/piper.yaml@master
   name: voice-kk-iseke-x-low
diff --git a/gallery/vllm.yaml b/gallery/vllm.yaml
new file mode 100644
index 00000000..5a2f16ce
--- /dev/null
+++ b/gallery/vllm.yaml
@@ -0,0 +1,32 @@
+---
+name: "vllm"
+
+config_file: |
+    context_size: 8192
+    parameters:
+      max_tokens: 8192
+    backend: vllm
+    function:
+      disable_no_action: true
+      grammar:
+        disable: true
+        parallel_calls: true
+        expect_strings_after_json: true
+    template:
+      use_tokenizer_template: true
+    # Uncomment to specify a quantization method (optional)
+    # quantization: "awq"
+    # Uncomment to limit the GPU memory utilization (vLLM default is 0.9 for 90%)
+    # gpu_memory_utilization: 0.5
+    # Uncomment to trust remote code from huggingface
+    # trust_remote_code: true
+    # Uncomment to enable eager execution
+    # enforce_eager: true
+    # Uncomment to specify the size of the CPU swap space per GPU (in GiB)
+    # swap_space: 2
+    # Uncomment to specify the maximum length of a sequence (including prompt and output)
+    # max_model_len: 32768
+    # Uncomment and specify the number of Tensor divisions.
+    # Allows you to partition and run large models. Performance gains are limited.
+    # https://github.com/vllm-project/vllm/issues/1435
+    # tensor_parallel_size: 2
diff --git a/go.mod b/go.mod
index 2d1d3de2..a3359abf 100644
--- a/go.mod
+++ b/go.mod
@@ -9,8 +9,10 @@ require (
 	github.com/M0Rf30/go-tiny-dream v0.0.0-20240425104733-c04fa463ace9
 	github.com/Masterminds/sprig/v3 v3.2.3
 	github.com/alecthomas/kong v0.9.0
+	github.com/census-instrumentation/opencensus-proto v0.4.1
 	github.com/charmbracelet/glamour v0.7.0
 	github.com/chasefleming/elem-go v0.26.0
+	github.com/cncf/xds/go v0.0.0-20240423153145-555b57ec207b
 	github.com/containerd/containerd v1.7.19
 	github.com/donomii/go-rwkv.cpp v0.0.0-20240228065144-661e7ae26d44
 	github.com/elliotchance/orderedmap/v2 v2.2.0
@@ -22,8 +24,11 @@ require (
 	github.com/gofiber/fiber/v2 v2.52.5
 	github.com/gofiber/swagger v1.0.0
 	github.com/gofiber/template/html/v2 v2.1.2
+	github.com/gofrs/flock v0.12.1
+	github.com/golang/protobuf v1.5.4
 	github.com/google/go-containerregistry v0.19.2
 	github.com/google/uuid v1.6.0
+	github.com/grpc-ecosystem/grpc-gateway v1.5.0
 	github.com/hpcloud/tail v1.0.0
 	github.com/ipfs/go-log v1.0.5
 	github.com/jaypipes/ghw v0.12.0
@@ -32,11 +37,10 @@ require (
 	github.com/libp2p/go-libp2p v0.36.2
 	github.com/mholt/archiver/v3 v3.5.1
 	github.com/microcosm-cc/bluemonday v1.0.26
-	github.com/mudler/edgevpn v0.27.2
-	github.com/mudler/go-processmanager v0.0.0-20230818213616-f204007f963c
+	github.com/mudler/edgevpn v0.28.3
+	github.com/mudler/go-processmanager v0.0.0-20240820160718-8b802d3ecf82
 	github.com/mudler/go-stable-diffusion v0.0.0-20240429204715-4a3cd6aeae6f
-	github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20240606155928-41c9013fa46a
-	github.com/onsi/ginkgo/v2 v2.20.0
+	github.com/onsi/ginkgo/v2 v2.20.1
 	github.com/onsi/gomega v1.34.1
 	github.com/ory/dockertest/v3 v3.10.0
 	github.com/otiai10/openaigo v1.7.0
@@ -56,6 +60,7 @@ require (
 	go.opentelemetry.io/otel/exporters/prometheus v0.50.0
 	go.opentelemetry.io/otel/metric v1.28.0
 	go.opentelemetry.io/otel/sdk/metric v1.28.0
+	google.golang.org/api v0.180.0
 	google.golang.org/grpc v1.65.0
 	google.golang.org/protobuf v1.34.2
 	gopkg.in/yaml.v2 v2.4.0
@@ -64,10 +69,19 @@ require (
 )
 
 require (
+	cel.dev/expr v0.15.0 // indirect
+	cloud.google.com/go/auth v0.4.1 // indirect
+	cloud.google.com/go/auth/oauth2adapt v0.2.2 // indirect
+	cloud.google.com/go/compute/metadata v0.3.0 // indirect
 	github.com/cpuguy83/go-md2man/v2 v2.0.4 // indirect
+	github.com/dave-gray101/v2keyauth v0.0.0-20240624150259-c45d584d25e2 // indirect
+	github.com/envoyproxy/protoc-gen-validate v1.0.4 // indirect
+	github.com/felixge/httpsnoop v1.0.4 // indirect
 	github.com/go-task/slim-sprig/v3 v3.0.0 // indirect
 	github.com/go-viper/mapstructure/v2 v2.0.0 // indirect
-	github.com/gofrs/flock v0.12.1 // indirect
+	github.com/google/s2a-go v0.1.7 // indirect
+	github.com/googleapis/enterprise-certificate-proxy v0.3.2 // indirect
+	github.com/googleapis/gax-go/v2 v2.12.4 // indirect
 	github.com/labstack/echo/v4 v4.12.0 // indirect
 	github.com/labstack/gommon v0.4.2 // indirect
 	github.com/moby/docker-image-spec v1.3.1 // indirect
@@ -89,11 +103,15 @@ require (
 	github.com/pion/turn/v2 v2.1.6 // indirect
 	github.com/pion/webrtc/v3 v3.3.0 // indirect
 	github.com/russross/blackfriday/v2 v2.1.0 // indirect
+	github.com/shirou/gopsutil/v4 v4.24.7 // indirect
 	github.com/urfave/cli/v2 v2.27.4 // indirect
 	github.com/valyala/fasttemplate v1.2.2 // indirect
 	github.com/wlynxg/anet v0.0.4 // indirect
 	github.com/xrash/smetrics v0.0.0-20240521201337-686a1a2994c1 // indirect
+	go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.52.0 // indirect
 	go.uber.org/mock v0.4.0 // indirect
+	golang.org/x/oauth2 v0.21.0 // indirect
+	google.golang.org/genproto/googleapis/api v0.0.0-20240617180043-68d350f18fd4 // indirect
 )
 
 require (
@@ -140,7 +158,7 @@ require (
 	github.com/go-audio/riff v1.0.0 // indirect
 	github.com/go-logr/logr v1.4.2 // indirect
 	github.com/go-logr/stdr v1.2.2 // indirect
-	github.com/go-ole/go-ole v1.2.6 // indirect
+	github.com/go-ole/go-ole v1.3.0 // indirect
 	github.com/go-openapi/jsonpointer v0.21.0 // indirect
 	github.com/go-openapi/jsonreference v0.21.0 // indirect
 	github.com/go-openapi/spec v0.21.0 // indirect
@@ -167,11 +185,11 @@ require (
 	github.com/huandu/xstrings v1.3.3 // indirect
 	github.com/huin/goupnp v1.3.0 // indirect
 	github.com/imdario/mergo v0.3.16 // indirect
-	github.com/ipfs/boxo v0.10.0 // indirect
+	github.com/ipfs/boxo v0.21.0 // indirect
 	github.com/ipfs/go-cid v0.4.1 // indirect
 	github.com/ipfs/go-datastore v0.6.0 // indirect
 	github.com/ipfs/go-log/v2 v2.5.1 // indirect
-	github.com/ipld/go-ipld-prime v0.20.0 // indirect
+	github.com/ipld/go-ipld-prime v0.21.0 // indirect
 	github.com/jackpal/go-nat-pmp v1.0.2 // indirect
 	github.com/jaypipes/pcidb v1.0.0 // indirect
 	github.com/jbenet/go-temp-err-catcher v0.1.0 // indirect
@@ -184,11 +202,11 @@ require (
 	github.com/libp2p/go-cidranger v1.1.0 // indirect
 	github.com/libp2p/go-flow-metrics v0.1.0 // indirect
 	github.com/libp2p/go-libp2p-asn-util v0.4.1 // indirect
-	github.com/libp2p/go-libp2p-kad-dht v0.25.2 // indirect
+	github.com/libp2p/go-libp2p-kad-dht v0.26.1 // indirect
 	github.com/libp2p/go-libp2p-kbucket v0.6.3 // indirect
 	github.com/libp2p/go-libp2p-pubsub v0.12.0 // indirect
 	github.com/libp2p/go-libp2p-record v0.2.0 // indirect
-	github.com/libp2p/go-libp2p-routing-helpers v0.7.2 // indirect
+	github.com/libp2p/go-libp2p-routing-helpers v0.7.4 // indirect
 	github.com/libp2p/go-msgio v0.3.0 // indirect
 	github.com/libp2p/go-nat v0.2.0 // indirect
 	github.com/libp2p/go-netroute v0.2.1 // indirect
@@ -196,7 +214,7 @@ require (
 	github.com/libp2p/go-yamux/v4 v4.0.1 // indirect
 	github.com/libp2p/zeroconf/v2 v2.2.0 // indirect
 	github.com/lucasb-eyer/go-colorful v1.2.0 // indirect
-	github.com/lufia/plan9stats v0.0.0-20211012122336-39d0f177ccd0 // indirect
+	github.com/lufia/plan9stats v0.0.0-20240819163618-b1d8f4d146e7 // indirect
 	github.com/mailru/easyjson v0.7.7 // indirect
 	github.com/marten-seemann/tcp v0.0.0-20210406111302-dfbc87cc63fd // indirect
 	github.com/mattn/go-colorable v0.1.13 // indirect
@@ -242,7 +260,7 @@ require (
 	github.com/pkoukk/tiktoken-go v0.1.6 // indirect
 	github.com/pmezard/go-difflib v1.0.0 // indirect
 	github.com/polydawn/refmt v0.89.0 // indirect
-	github.com/power-devops/perfstat v0.0.0-20210106213030-5aafc221ea8c // indirect
+	github.com/power-devops/perfstat v0.0.0-20240221224432-82ca36839d55 // indirect
 	github.com/prometheus/client_model v0.6.1 // indirect
 	github.com/prometheus/common v0.55.0 // indirect
 	github.com/prometheus/procfs v0.15.1 // indirect
@@ -260,14 +278,14 @@ require (
 	github.com/spf13/cast v1.5.0 // indirect
 	github.com/swaggo/files/v2 v2.0.0 // indirect
 	github.com/tinylib/msgp v1.1.8 // indirect
-	github.com/tklauser/go-sysconf v0.3.12 // indirect
-	github.com/tklauser/numcpus v0.6.1 // indirect
+	github.com/tklauser/go-sysconf v0.3.14 // indirect
+	github.com/tklauser/numcpus v0.8.0 // indirect
 	github.com/ulikunitz/xz v0.5.9 // indirect
 	github.com/valyala/bytebufferpool v1.0.0 // indirect
 	github.com/valyala/tcplisten v1.0.0 // indirect
 	github.com/vbatts/tar-split v0.11.3 // indirect
-	github.com/vishvananda/netlink v1.2.1-beta.2 // indirect
-	github.com/vishvananda/netns v0.0.0-20211101163701-50045581ed74 // indirect
+	github.com/vishvananda/netlink v1.3.0 // indirect
+	github.com/vishvananda/netns v0.0.4 // indirect
 	github.com/whyrusleeping/go-keyspace v0.0.0-20160322163242-5b898ac5add1 // indirect
 	github.com/xeipuuv/gojsonpointer v0.0.0-20190905194746-02993c407bfb // indirect
 	github.com/xeipuuv/gojsonreference v0.0.0-20180127040603-bd5ef7bd5415 // indirect
@@ -295,8 +313,8 @@ require (
 	golang.zx2c4.com/wintun v0.0.0-20211104114900-415007cec224 // indirect
 	golang.zx2c4.com/wireguard v0.0.0-20220703234212-c31a7b1ab478 // indirect
 	golang.zx2c4.com/wireguard/windows v0.5.3 // indirect
-	gonum.org/v1/gonum v0.13.0 // indirect
-	google.golang.org/genproto/googleapis/rpc v0.0.0-20240528184218-531527333157 // indirect
+	gonum.org/v1/gonum v0.15.0 // indirect
+	google.golang.org/genproto/googleapis/rpc v0.0.0-20240617180043-68d350f18fd4 // indirect
 	gopkg.in/fsnotify.v1 v1.4.7 // indirect
 	gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7 // indirect
 	howett.net/plist v1.0.0 // indirect
diff --git a/go.sum b/go.sum
index 962bb94b..1dd44a5b 100644
--- a/go.sum
+++ b/go.sum
@@ -1,7 +1,15 @@
+cel.dev/expr v0.15.0 h1:O1jzfJCQBfL5BFoYktaxwIhuttaQPsVWerH9/EEKx0w=
+cel.dev/expr v0.15.0/go.mod h1:TRSuuV7DlVCE/uwv5QbAiW/v8l5O8C4eEPHeu7gf7Sg=
 cloud.google.com/go v0.26.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
 cloud.google.com/go v0.31.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
 cloud.google.com/go v0.34.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
 cloud.google.com/go v0.37.0/go.mod h1:TS1dMSSfndXH133OKGwekG838Om/cQT0BUHV3HcBgoo=
+cloud.google.com/go/auth v0.4.1 h1:Z7YNIhlWRtrnKlZke7z3GMqzvuYzdc2z98F9D1NV5Hg=
+cloud.google.com/go/auth v0.4.1/go.mod h1:QVBuVEKpCn4Zp58hzRGvL0tjRGU0YqdRTdCHM1IHnro=
+cloud.google.com/go/auth/oauth2adapt v0.2.2 h1:+TTV8aXpjeChS9M+aTtN/TjdQnzJvmzKFt//oWu7HX4=
+cloud.google.com/go/auth/oauth2adapt v0.2.2/go.mod h1:wcYjgpZI9+Yu7LyYBg4pqSiaRkfEK3GQcpb7C/uyF1Q=
+cloud.google.com/go/compute/metadata v0.3.0 h1:Tz+eQXMEqDIKRsmY3cHTL6FVaynIjX2QxYC4trgAKZc=
+cloud.google.com/go/compute/metadata v0.3.0/go.mod h1:zFmK7XCadkQkj6TtorcaGlCW1hT1fIilQDwofLpJ20k=
 dario.cat/mergo v1.0.0 h1:AGCNq9Evsj31mOgNPcLyXc+4PNABt905YmuqPYYpBWk=
 dario.cat/mergo v1.0.0/go.mod h1:uNxQE+84aUszobStD9th8a29P2fMDhsBdgRYvZOxGmk=
 dmitri.shuralyov.com/app/changes v0.0.0-20180602232624-0a106ad413e3/go.mod h1:Yl+fi1br7+Rr3LqpNJf1/uxUdtRUV+Tnj0o93V2B9MU=
@@ -61,6 +69,8 @@ github.com/c-robinson/iplib v1.0.8/go.mod h1:i3LuuFL1hRT5gFpBRnEydzw8R6yhGkF4szN
 github.com/cenkalti/backoff/v4 v4.3.0 h1:MyRJ/UdXutAwSAT+s3wNd7MfTIcy71VQueUuFK343L8=
 github.com/cenkalti/backoff/v4 v4.3.0/go.mod h1:Y3VNntkOUPxTVeUxJ/G5vcM//AlwfmyYozVcomhLiZE=
 github.com/census-instrumentation/opencensus-proto v0.2.1/go.mod h1:f6KPmirojxKA12rnyqOA5BBL4O983OfeGPqjHWSTneU=
+github.com/census-instrumentation/opencensus-proto v0.4.1 h1:iKLQ0xPNFxR/2hzXZMrBo8f1j86j5WHzznCCQxV/b8g=
+github.com/census-instrumentation/opencensus-proto v0.4.1/go.mod h1:4T9NM4+4Vw91VeyqjLS6ao50K5bOcLKN6Q42XnYaRYw=
 github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
 github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
 github.com/charmbracelet/glamour v0.7.0 h1:2BtKGZ4iVJCDfMF229EzbeR1QRKLWztO9dMtjmqZSng=
@@ -70,6 +80,8 @@ github.com/chasefleming/elem-go v0.26.0/go.mod h1:hz73qILBIKnTgOujnSMtEj20/epI+f
 github.com/cilium/ebpf v0.2.0/go.mod h1:To2CFviqOWL/M0gIMsvSMlqe7em/l1ALkX1PyjrX2Qs=
 github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
 github.com/cncf/udpa/go v0.0.0-20191209042840-269d4d468f6f/go.mod h1:M8M6+tZqaGXZJjfX53e64911xZQV5JYwmTeXPW+k8Sc=
+github.com/cncf/xds/go v0.0.0-20240423153145-555b57ec207b h1:ga8SEFjZ60pxLcmhnThWgvH2wg8376yUJmPhEH4H3kw=
+github.com/cncf/xds/go v0.0.0-20240423153145-555b57ec207b/go.mod h1:W+zGtBO5Y1IgJhy4+A9GOqVhqLpfZi+vwmdNXUehLA8=
 github.com/containerd/cgroups v0.0.0-20201119153540-4cbc285b3327/go.mod h1:ZJeTFisyysqgcCdecO57Dj79RfL0LNeGiFUqLYQRYLE=
 github.com/containerd/cgroups v1.1.0 h1:v8rEWFl6EoqHB+swVNjVoCJE8o3jX7e8nqBGPLaDFBM=
 github.com/containerd/cgroups v1.1.0/go.mod h1:6ppBcbh/NOOUU+dMKrykgaBnK9lCIBxHqJDGwsa1mIw=
@@ -98,6 +110,8 @@ github.com/creachadair/otp v0.4.2 h1:ngNMaD6Tzd7UUNRFyed7ykZFn/Wr5sSs5ffqZWm9pu8
 github.com/creachadair/otp v0.4.2/go.mod h1:DqV9hJyUbcUme0pooYfiFvvMe72Aua5sfhNzwfZvk40=
 github.com/creack/pty v1.1.18 h1:n56/Zwd5o6whRC5PMGretI4IdRLlmBXYNjScPaBgsbY=
 github.com/creack/pty v1.1.18/go.mod h1:MOBLtS5ELjhRRrroQr9kyvTxUAFNvYEK993ew/Vr4O4=
+github.com/dave-gray101/v2keyauth v0.0.0-20240624150259-c45d584d25e2 h1:flLYmnQFZNo04x2NPehMbf30m7Pli57xwZ0NFqR/hb0=
+github.com/dave-gray101/v2keyauth v0.0.0-20240624150259-c45d584d25e2/go.mod h1:NtWqRzAp/1tw+twkW8uuBenEVVYndEAZACWU3F3xdoQ=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
@@ -129,8 +143,6 @@ github.com/dsnet/compress v0.0.2-0.20210315054119-f66993602bf5/go.mod h1:qssHWj6
 github.com/dsnet/golib v0.0.0-20171103203638-1ea166775780/go.mod h1:Lj+Z9rebOhdfkVLjJ8T6VcRQv3SXugXy999NBtR9aFY=
 github.com/dustin/go-humanize v1.0.0/go.mod h1:HtrtbFcZ19U5GC7JDqmcUSB87Iq5E25KnS6fMYU6eOk=
 github.com/elastic/gosigar v0.12.0/go.mod h1:iXRIGg2tLnu7LBdpqzyQfGDEidKCfWcCMS0WKyPWoMs=
-github.com/elastic/gosigar v0.14.2 h1:Dg80n8cr90OZ7x+bAax/QjoW/XqTI11RmA79ZwIm9/4=
-github.com/elastic/gosigar v0.14.2/go.mod h1:iXRIGg2tLnu7LBdpqzyQfGDEidKCfWcCMS0WKyPWoMs=
 github.com/elastic/gosigar v0.14.3 h1:xwkKwPia+hSfg9GqrCUKYdId102m9qTJIIr7egmK/uo=
 github.com/elastic/gosigar v0.14.3/go.mod h1:iXRIGg2tLnu7LBdpqzyQfGDEidKCfWcCMS0WKyPWoMs=
 github.com/elliotchance/orderedmap/v2 v2.2.0 h1:7/2iwO98kYT4XkOjA9mBEIwvi4KpGB4cyHeOFOnj4Vk=
@@ -139,15 +151,18 @@ github.com/envoyproxy/go-control-plane v0.9.0/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymF
 github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
 github.com/envoyproxy/go-control-plane v0.9.4/go.mod h1:6rpuAdCZL397s3pYoYcLgu1mIlRU8Am5FuJP05cCM98=
 github.com/envoyproxy/protoc-gen-validate v0.1.0/go.mod h1:iSmxcyjqTsJpI2R4NaDN7+kN2VEUnK/pcBlmesArF7c=
+github.com/envoyproxy/protoc-gen-validate v1.0.4 h1:gVPz/FMfvh57HdSJQyvBtF00j8JU4zdyUgIUNhlgg0A=
+github.com/envoyproxy/protoc-gen-validate v1.0.4/go.mod h1:qys6tmnRsYrQqIhm2bvKZH4Blx/1gTIZ2UKVY1M+Yew=
+github.com/felixge/httpsnoop v1.0.4 h1:NFTV2Zj1bL4mc9sqWACXbQFVBBg2W3GPvqp8/ESS2Wg=
+github.com/felixge/httpsnoop v1.0.4/go.mod h1:m8KPJKqk1gH5J9DgRY2ASl2lWCfGKXixSwevea8zH2U=
 github.com/flynn/go-shlex v0.0.0-20150515145356-3f9db97f8568/go.mod h1:xEzjJPgXI435gkrCt3MPfRiAkVrwSbHsst4LCFVfpJc=
 github.com/flynn/noise v1.1.0 h1:KjPQoQCEFdZDiP03phOvGi11+SVVhBG2wOWAorLsstg=
 github.com/flynn/noise v1.1.0/go.mod h1:xbMo+0i6+IGbYdJhF31t2eR1BIU0CYc12+BNAKwUTag=
 github.com/francoispqt/gojay v1.2.13 h1:d2m3sFjloqoIUQU3TsHBgj6qg/BVGlTBeHDUmyJnXKk=
 github.com/francoispqt/gojay v1.2.13/go.mod h1:ehT5mTG4ua4581f1++1WLG0vPdaA9HaiDsoyrBGkyDY=
-github.com/frankban/quicktest v1.14.4 h1:g2rn0vABPOOXmZUj+vbmUp0lPoXEMuhTpIluN0XL9UY=
-github.com/frankban/quicktest v1.14.4/go.mod h1:4ptaffx2x8+WTWXmUCuVU6aPUX1/Mz7zb5vbUoiM6w0=
+github.com/frankban/quicktest v1.14.6 h1:7Xjx+VpznH+oBnejlPUj8oUpdxnVs4f8XU8WnHkI4W8=
+github.com/frankban/quicktest v1.14.6/go.mod h1:4ptaffx2x8+WTWXmUCuVU6aPUX1/Mz7zb5vbUoiM6w0=
 github.com/fsnotify/fsnotify v1.4.7/go.mod h1:jwhsz4b93w/PPRr/qN1Yymfu8t87LnFCMoQvtojpjFo=
-github.com/fsnotify/fsnotify v1.4.9/go.mod h1:znqG4EE+3YCdAaPaxE2ZRY/06pZUdp0tY4IgpuI1SZQ=
 github.com/fsnotify/fsnotify v1.7.0 h1:8JEhPFa5W2WU7YfeZzPNqzMP6Lwt7L2715Ggo0nosvA=
 github.com/fsnotify/fsnotify v1.7.0/go.mod h1:40Bi/Hjc2AVfZrqy+aj+yEI+/bRxZnMJyTJwOpGvigM=
 github.com/ggerganov/whisper.cpp/bindings/go v0.0.0-20240626202019-c118733a29ad h1:dQ93Vd6i25o+zH9vvnZ8mu7jtJQ6jT3D+zE3V8Q49n0=
@@ -168,8 +183,9 @@ github.com/go-logr/logr v1.4.2/go.mod h1:9T104GzyrTigFIr8wt5mBrctHMim0Nb2HLGrmQ4
 github.com/go-logr/stdr v1.2.2 h1:hSWxHoqTgW2S2qGc0LTAI563KZ5YKYRhT3MFKZMbjag=
 github.com/go-logr/stdr v1.2.2/go.mod h1:mMo/vtBO5dYbehREoey6XUKy/eSumjCCveDpRre4VKE=
 github.com/go-ole/go-ole v1.2.5/go.mod h1:pprOEPIfldk/42T2oK7lQ4v4JSDwmV0As9GaiUsvbm0=
-github.com/go-ole/go-ole v1.2.6 h1:/Fpf6oFPoeFik9ty7siob0G6Ke8QvQEuVcuChpwXzpY=
 github.com/go-ole/go-ole v1.2.6/go.mod h1:pprOEPIfldk/42T2oK7lQ4v4JSDwmV0As9GaiUsvbm0=
+github.com/go-ole/go-ole v1.3.0 h1:Dt6ye7+vXGIKZ7Xtk4s6/xVdGDQynvom7xCFEdWr6uE=
+github.com/go-ole/go-ole v1.3.0/go.mod h1:5LS6F96DhAwUc7C+1HLexzMXY1xGRSryjyPPKW6zv78=
 github.com/go-openapi/jsonpointer v0.21.0 h1:YgdVicSA9vH5RiHs9TZW5oyafXZFc6+2Vc1rr/O9oNQ=
 github.com/go-openapi/jsonpointer v0.21.0/go.mod h1:IUyH9l/+uyhIYQ/PXVA41Rexl+kOkAPDdXEYns6fzUY=
 github.com/go-openapi/jsonreference v0.21.0 h1:Rs+Y7hSXT83Jacb7kFyjn4ijOuVGSvOdF2+tg1TRrwQ=
@@ -184,7 +200,6 @@ github.com/go-skynet/go-llama.cpp v0.0.0-20240314183750-6a8041ef6b46 h1:lALhXzDk
 github.com/go-skynet/go-llama.cpp v0.0.0-20240314183750-6a8041ef6b46/go.mod h1:iub0ugfTnflE3rcIuqV2pQSo15nEw3GLW/utm5gyERo=
 github.com/go-sql-driver/mysql v1.7.1 h1:lUIinVbN1DY0xBg0eMOzmmtGoHwWBbvnWubQUrtU8EI=
 github.com/go-sql-driver/mysql v1.7.1/go.mod h1:OXbVy3sEdcQ2Doequ6Z5BW6fXNQTmx+9S1MCJN5yJMI=
-github.com/go-task/slim-sprig v0.0.0-20210107165309-348f09dbbbc0/go.mod h1:fyg7847qk6SyHyPtNmDHnmrv/HOrqktSC+C9fM+CJOE=
 github.com/go-task/slim-sprig/v3 v3.0.0 h1:sUs3vkvUymDpBKi3qH1YSqBQk9+9D/8M2mN1vB6EwHI=
 github.com/go-task/slim-sprig/v3 v3.0.0/go.mod h1:W848ghGpv3Qj3dhTPRyJypKRiqCdHZiAzKg9hl15HA8=
 github.com/go-viper/mapstructure/v2 v2.0.0 h1:dhn8MZ1gZ0mzeodTG3jt5Vj/o87xZKuNAprG2mQfMfc=
@@ -228,10 +243,9 @@ github.com/golang/protobuf v1.4.0-rc.2/go.mod h1:LlEzMj4AhA7rCAGe4KMBDvJI+AwstrU
 github.com/golang/protobuf v1.4.0-rc.4.0.20200313231945-b860323f09d0/go.mod h1:WU3c8KckQ9AFe+yFwt9sWVRKCVIyN9cPHBJSNnbL67w=
 github.com/golang/protobuf v1.4.0/go.mod h1:jodUvKwWbYaEsadDk5Fwe5c77LiNKVO9IDvqG2KuDX0=
 github.com/golang/protobuf v1.4.1/go.mod h1:U8fpvMrcmy5pZrNK1lt4xCsGvpyWQ/VVv6QDs8UjoX8=
-github.com/golang/protobuf v1.4.2/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
 github.com/golang/protobuf v1.4.3/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
-github.com/golang/protobuf v1.5.0/go.mod h1:FsONVRAS9T7sI+LIUmWTfcYkHO4aIWwzhcaSAoJOfIk=
-github.com/golang/protobuf v1.5.2/go.mod h1:XVQd3VNwM+JqD3oG2Ue2ip4fOMUkwXdXDdiuN0vRsmY=
+github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek=
+github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps=
 github.com/golang/snappy v0.0.2/go.mod h1:/XxbfmMg8lxefKM7IXC3fBNl/7bRcc72aCRzEWrmP2Q=
 github.com/golang/snappy v0.0.4 h1:yAGX7huGHXlcLOEtBnF4w7FQwA26wojNCwOYAEhLjQM=
 github.com/golang/snappy v0.0.4/go.mod h1:/XxbfmMg8lxefKM7IXC3fBNl/7bRcc72aCRzEWrmP2Q=
@@ -246,8 +260,6 @@ github.com/google/go-cmp v0.5.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/
 github.com/google/go-cmp v0.5.2/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
 github.com/google/go-cmp v0.5.3/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
 github.com/google/go-cmp v0.5.5/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
-github.com/google/go-cmp v0.5.6/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
-github.com/google/go-cmp v0.5.9/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
 github.com/google/go-cmp v0.6.0 h1:ofyhxvXcZhMsU5ulbFiLKl/XBFqE1GSq7atu8tAmTRI=
 github.com/google/go-cmp v0.6.0/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
 github.com/google/go-containerregistry v0.19.2 h1:TannFKE1QSajsP6hPWb5oJNgKe1IKjHukIKDUmvsV6w=
@@ -258,11 +270,11 @@ github.com/google/gopacket v1.1.19 h1:ves8RnFZPGiFnTS0uPQStjwru6uO6h+nlr9j6fL7kF
 github.com/google/gopacket v1.1.19/go.mod h1:iJ8V8n6KS+z2U1A8pUwu8bW5SyEMkXJB8Yo/Vo+TKTo=
 github.com/google/martian v2.1.0+incompatible/go.mod h1:9I4somxYTbIHy5NJKHRl3wXiIaQGbYVAs8BPL6v8lEs=
 github.com/google/pprof v0.0.0-20181206194817-3ea8567a2e57/go.mod h1:zfwlbNMJ+OItoe0UupaVj+oy1omPYYDuagoSzA8v9mc=
-github.com/google/pprof v0.0.0-20240424215950-a892ee059fd6 h1:k7nVchz72niMH6YLQNvHSdIE7iqsQxK1P41mySCvssg=
-github.com/google/pprof v0.0.0-20240424215950-a892ee059fd6/go.mod h1:kf6iHlnVGwgKolg33glAes7Yg/8iWP8ukqeldJSO7jw=
 github.com/google/pprof v0.0.0-20240727154555-813a5fbdbec8 h1:FKHo8hFI3A+7w0aUQuYXQ+6EN5stWmeY/AZqtM8xk9k=
 github.com/google/pprof v0.0.0-20240727154555-813a5fbdbec8/go.mod h1:K1liHPHnj73Fdn/EKuT8nrFqBihUSKXoLYU0BuatOYo=
 github.com/google/renameio v0.1.0/go.mod h1:KWCgfxg9yswjAJkECMjeO8J8rahYeXnNhOm40UhjYkI=
+github.com/google/s2a-go v0.1.7 h1:60BLSyTrOV4/haCDW4zb1guZItoSq8foHCXrAnjBo/o=
+github.com/google/s2a-go v0.1.7/go.mod h1:50CgR4k1jNlWBu4UfS4AcfhVe1r6pdZPygJ3R8F0Qdw=
 github.com/google/shlex v0.0.0-20191202100458-e7afc7fbc510 h1:El6M4kTTCOh6aBiKaUGG7oYTSPP8MxqL4YI3kZKwcP4=
 github.com/google/shlex v0.0.0-20191202100458-e7afc7fbc510/go.mod h1:pupxD2MaaD3pAXIBCelhxNneeOaAeabZDe5s4K6zSpQ=
 github.com/google/uuid v1.1.1/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
@@ -270,8 +282,12 @@ github.com/google/uuid v1.1.2/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+
 github.com/google/uuid v1.3.1/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
+github.com/googleapis/enterprise-certificate-proxy v0.3.2 h1:Vie5ybvEvT75RniqhfFxPRy3Bf7vr3h0cechB90XaQs=
+github.com/googleapis/enterprise-certificate-proxy v0.3.2/go.mod h1:VLSiSSBs/ksPL8kq3OBOQ6WRI2QnaFynd1DCjZ62+V0=
 github.com/googleapis/gax-go v2.0.0+incompatible/go.mod h1:SFVmujtThgffbyetf+mdk2eWhX2bMyUtNHzFKcPA9HY=
 github.com/googleapis/gax-go/v2 v2.0.3/go.mod h1:LLvjysVCY1JZeum8Z6l8qUty8fiNwE08qbEPm1M08qg=
+github.com/googleapis/gax-go/v2 v2.12.4 h1:9gWcmF85Wvq4ryPFvGFaOgPIs1AQX0d0bcbGw4Z96qg=
+github.com/googleapis/gax-go/v2 v2.12.4/go.mod h1:KYEYLorsnIGDi/rPC8b5TdlB9kbKoFubselGIoBMCwI=
 github.com/gopherjs/gopherjs v0.0.0-20181017120253-0766667cb4d1/go.mod h1:wJfORRmW1u3UXTncJ5qlYoELFm8eSnnEO6hX4iZ3EWY=
 github.com/gopherjs/gopherjs v0.0.0-20190430165422-3e4dfb77656c h1:7lF+Vz0LqiRidnzC1Oq86fpX1q/iEv2KJdrCtttYjT4=
 github.com/gopherjs/gopherjs v0.0.0-20190430165422-3e4dfb77656c/go.mod h1:wJfORRmW1u3UXTncJ5qlYoELFm8eSnnEO6hX4iZ3EWY=
@@ -280,6 +296,7 @@ github.com/gorilla/css v1.0.1/go.mod h1:BvnYkspnSzMmwRK+b8/xgNPLiIuNZr6vbZBTPQ2A
 github.com/gorilla/websocket v1.5.3 h1:saDtZ6Pbx/0u+bgYQ3q96pZgCzfhKXGPqt7kZ72aNNg=
 github.com/gorilla/websocket v1.5.3/go.mod h1:YR8l580nyteQvAITg2hZ9XVh4b55+EU/adAjf1fMHhE=
 github.com/gregjones/httpcache v0.0.0-20180305231024-9cad4c3443a7/go.mod h1:FecbI9+v66THATjSRHfNgh1IVFe/9kFxbXtjV0ctIMA=
+github.com/grpc-ecosystem/grpc-gateway v1.5.0 h1:WcmKMm43DR7RdtlkEXQJyo5ws8iTp98CyhCCbOHMvNI=
 github.com/grpc-ecosystem/grpc-gateway v1.5.0/go.mod h1:RSKVYQBd5MCa4OVpNdGskqpgL2+G+NZTnrVHpWWfpdw=
 github.com/hashicorp/errwrap v1.0.0/go.mod h1:YH+1FKiLXxHSkmPseP+kNlulaMuP3n2brvKWEqk/Jc4=
 github.com/hashicorp/errwrap v1.1.0 h1:OxrOeh75EUXMY8TBjag2fzXGZ40LB6IKw45YeGUDY2I=
@@ -303,23 +320,23 @@ github.com/huin/goupnp v1.3.0/go.mod h1:gnGPsThkYa7bFi/KWmEysQRf48l2dvR5bxr2OFck
 github.com/imdario/mergo v0.3.11/go.mod h1:jmQim1M+e3UYxmgPu/WyfjB3N3VflVyUjjjwH0dnCYA=
 github.com/imdario/mergo v0.3.16 h1:wwQJbIsHYGMUyLSPrEq1CT16AhnhNJQ51+4fdHUnCl4=
 github.com/imdario/mergo v0.3.16/go.mod h1:WBLT9ZmE3lPoWsEzCh9LPo3TiwVN+ZKEjmz+hD27ysY=
-github.com/ipfs/boxo v0.10.0 h1:tdDAxq8jrsbRkYoF+5Rcqyeb91hgWe2hp7iLu7ORZLY=
-github.com/ipfs/boxo v0.10.0/go.mod h1:Fg+BnfxZ0RPzR0nOodzdIq3A7KgoWAOWsEIImrIQdBM=
+github.com/ipfs/boxo v0.21.0 h1:XpGXb+TQQ0IUdYaeAxGzWjSs6ow/Lce148A/2IbRDVE=
+github.com/ipfs/boxo v0.21.0/go.mod h1:NmweAYeY1USOaJJxouy7DLr/Y5M8UBSsCI2KRivO+TY=
 github.com/ipfs/go-cid v0.4.1 h1:A/T3qGvxi4kpKWWcPC/PgbvDA2bjVLO7n4UeVwnbs/s=
 github.com/ipfs/go-cid v0.4.1/go.mod h1:uQHwDeX4c6CtyrFwdqyhpNcxVewur1M7l7fNU7LKwZk=
 github.com/ipfs/go-datastore v0.6.0 h1:JKyz+Gvz1QEZw0LsX1IBn+JFCJQH4SJVFtM4uWU0Myk=
 github.com/ipfs/go-datastore v0.6.0/go.mod h1:rt5M3nNbSO/8q1t4LNkLyUwRs8HupMeN/8O4Vn9YAT8=
 github.com/ipfs/go-detect-race v0.0.1 h1:qX/xay2W3E4Q1U7d9lNs1sU9nvguX0a7319XbyQ6cOk=
 github.com/ipfs/go-detect-race v0.0.1/go.mod h1:8BNT7shDZPo99Q74BpGMK+4D8Mn4j46UU0LZ723meps=
-github.com/ipfs/go-ipfs-util v0.0.2 h1:59Sswnk1MFaiq+VcaknX7aYEyGyGDAA73ilhEK2POp8=
-github.com/ipfs/go-ipfs-util v0.0.2/go.mod h1:CbPtkWJzjLdEcezDns2XYaehFVNXG9zrdrtMecczcsQ=
+github.com/ipfs/go-ipfs-util v0.0.3 h1:2RFdGez6bu2ZlZdI+rWfIdbQb1KudQp3VGwPtdNCmE0=
+github.com/ipfs/go-ipfs-util v0.0.3/go.mod h1:LHzG1a0Ig4G+iZ26UUOMjHd+lfM84LZCrn17xAKWBvs=
 github.com/ipfs/go-log v1.0.5 h1:2dOuUCB1Z7uoczMWgAyDck5JLb72zHzrMnGnCNNbvY8=
 github.com/ipfs/go-log v1.0.5/go.mod h1:j0b8ZoR+7+R99LD9jZ6+AJsrzkPbSXbZfGakb5JPtIo=
 github.com/ipfs/go-log/v2 v2.1.3/go.mod h1:/8d0SH3Su5Ooc31QlL1WysJhvyOTDCjcCZ9Axpmri6g=
 github.com/ipfs/go-log/v2 v2.5.1 h1:1XdUzF7048prq4aBjDQQ4SL5RxftpRGdXhNRwKSAlcY=
 github.com/ipfs/go-log/v2 v2.5.1/go.mod h1:prSpmC1Gpllc9UYWxDiZDreBYw7zp4Iqp1kOLU9U5UI=
-github.com/ipld/go-ipld-prime v0.20.0 h1:Ud3VwE9ClxpO2LkCYP7vWPc0Fo+dYdYzgxUJZ3uRG4g=
-github.com/ipld/go-ipld-prime v0.20.0/go.mod h1:PzqZ/ZR981eKbgdr3y2DJYeD/8bgMawdGVlJDE8kK+M=
+github.com/ipld/go-ipld-prime v0.21.0 h1:n4JmcpOlPDIxBcY037SVfpd1G+Sj1nKZah0m6QH9C2E=
+github.com/ipld/go-ipld-prime v0.21.0/go.mod h1:3RLqy//ERg/y5oShXXdx5YIp50cFGOanyMctpPjsvxQ=
 github.com/jackpal/go-nat-pmp v1.0.2 h1:KzKSgb7qkJvOUTqYl9/Hg/me3pWgBmERKrTGD7BdWus=
 github.com/jackpal/go-nat-pmp v1.0.2/go.mod h1:QPH045xvCAeXUZOxsnwmrtiCoxIr9eob+4orBN1SBKc=
 github.com/jaypipes/ghw v0.12.0 h1:xU2/MDJfWmBhJnujHY9qwXQLs3DBsf0/Xa9vECY0Tho=
@@ -365,6 +382,8 @@ github.com/kr/pty v1.1.3/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
 github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
 github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
 github.com/kr/text v0.2.0/go.mod h1:eLer722TekiGuMkidMxC/pM04lWEeraHUUmBw8l2grE=
+github.com/kylelemons/godebug v1.1.0 h1:RPNrshWIDI6G2gRW9EHilWtl7Z6Sb1BR0xunSBf0SNc=
+github.com/kylelemons/godebug v1.1.0/go.mod h1:9/0rRGxNHcop5bhtWyNeEfOS8JIWk580+fNqagV/RAw=
 github.com/labstack/echo/v4 v4.12.0 h1:IKpw49IMryVB2p1a4dzwlhP1O2Tf2E0Ir/450lH+kI0=
 github.com/labstack/echo/v4 v4.12.0/go.mod h1:UP9Cr2DJXbOK3Kr9ONYzNowSh7HP0aG0ShAyycHSJvM=
 github.com/labstack/gommon v0.4.2 h1:F8qTUNXgG1+6WQmqoUWnz8WiEU60mXVVw0P4ht1WRA0=
@@ -377,26 +396,20 @@ github.com/libp2p/go-cidranger v1.1.0 h1:ewPN8EZ0dd1LSnrtuwd4709PXVcITVeuwbag38y
 github.com/libp2p/go-cidranger v1.1.0/go.mod h1:KWZTfSr+r9qEo9OkI9/SIEeAtw+NNoU0dXIXt15Okic=
 github.com/libp2p/go-flow-metrics v0.1.0 h1:0iPhMI8PskQwzh57jB9WxIuIOQ0r+15PChFGkx3Q3WM=
 github.com/libp2p/go-flow-metrics v0.1.0/go.mod h1:4Xi8MX8wj5aWNDAZttg6UPmc0ZrnFNsMtpsYUClFtro=
-github.com/libp2p/go-libp2p v0.35.2 h1:287oHbuplkrLdAF+syB0n/qDgd50AUBtEODqS0e0HDs=
-github.com/libp2p/go-libp2p v0.35.2/go.mod h1:RKCDNt30IkFipGL0tl8wQW/3zVWEGFUZo8g2gAKxwjU=
-github.com/libp2p/go-libp2p v0.35.4 h1:FDiBUYLkueFwsuNJUZaxKRdpKvBOWU64qQPL768bSeg=
-github.com/libp2p/go-libp2p v0.35.4/go.mod h1:RKCDNt30IkFipGL0tl8wQW/3zVWEGFUZo8g2gAKxwjU=
 github.com/libp2p/go-libp2p v0.36.2 h1:BbqRkDaGC3/5xfaJakLV/BrpjlAuYqSB0lRvtzL3B/U=
 github.com/libp2p/go-libp2p v0.36.2/go.mod h1:XO3joasRE4Eup8yCTTP/+kX+g92mOgRaadk46LmPhHY=
 github.com/libp2p/go-libp2p-asn-util v0.4.1 h1:xqL7++IKD9TBFMgnLPZR6/6iYhawHKHl950SO9L6n94=
 github.com/libp2p/go-libp2p-asn-util v0.4.1/go.mod h1:d/NI6XZ9qxw67b4e+NgpQexCIiFYJjErASrYW4PFDN8=
-github.com/libp2p/go-libp2p-kad-dht v0.25.2 h1:FOIk9gHoe4YRWXTu8SY9Z1d0RILol0TrtApsMDPjAVQ=
-github.com/libp2p/go-libp2p-kad-dht v0.25.2/go.mod h1:6za56ncRHYXX4Nc2vn8z7CZK0P4QiMcrn77acKLM2Oo=
+github.com/libp2p/go-libp2p-kad-dht v0.26.1 h1:AazV3LCImYVkDUGAHx5lIEgZ9iUI2QQKH5GMRQU8uEA=
+github.com/libp2p/go-libp2p-kad-dht v0.26.1/go.mod h1:mqRUGJ/+7ziQ3XknU2kKHfsbbgb9xL65DXjPOJwmZF8=
 github.com/libp2p/go-libp2p-kbucket v0.6.3 h1:p507271wWzpy2f1XxPzCQG9NiN6R6lHL9GiSErbQQo0=
 github.com/libp2p/go-libp2p-kbucket v0.6.3/go.mod h1:RCseT7AH6eJWxxk2ol03xtP9pEHetYSPXOaJnOiD8i0=
-github.com/libp2p/go-libp2p-pubsub v0.11.0 h1:+JvS8Kty0OiyUiN0i8H5JbaCgjnJTRnTHe4rU88dLFc=
-github.com/libp2p/go-libp2p-pubsub v0.11.0/go.mod h1:QEb+hEV9WL9wCiUAnpY29FZR6W3zK8qYlaml8R4q6gQ=
 github.com/libp2p/go-libp2p-pubsub v0.12.0 h1:PENNZjSfk8KYxANRlpipdS7+BfLmOl3L2E/6vSNjbdI=
 github.com/libp2p/go-libp2p-pubsub v0.12.0/go.mod h1:Oi0zw9aw8/Y5GC99zt+Ef2gYAl+0nZlwdJonDyOz/sE=
 github.com/libp2p/go-libp2p-record v0.2.0 h1:oiNUOCWno2BFuxt3my4i1frNrt7PerzB3queqa1NkQ0=
 github.com/libp2p/go-libp2p-record v0.2.0/go.mod h1:I+3zMkvvg5m2OcSdoL0KPljyJyvNDFGKX7QdlpYUcwk=
-github.com/libp2p/go-libp2p-routing-helpers v0.7.2 h1:xJMFyhQ3Iuqnk9Q2dYE1eUTzsah7NLw3Qs2zjUV78T0=
-github.com/libp2p/go-libp2p-routing-helpers v0.7.2/go.mod h1:cN4mJAD/7zfPKXBcs9ze31JGYAZgzdABEm+q/hkswb8=
+github.com/libp2p/go-libp2p-routing-helpers v0.7.4 h1:6LqS1Bzn5CfDJ4tzvP9uwh42IB7TJLNFJA6dEeGBv84=
+github.com/libp2p/go-libp2p-routing-helpers v0.7.4/go.mod h1:we5WDj9tbolBXOuF1hGOkR+r7Uh1408tQbAKaT5n1LE=
 github.com/libp2p/go-libp2p-testing v0.12.0 h1:EPvBb4kKMWO29qP4mZGyhVzUyR25dvfUIK5WDu6iPUA=
 github.com/libp2p/go-libp2p-testing v0.12.0/go.mod h1:KcGDRXyN7sQCllucn1cOOS+Dmm7ujhfEyXQL5lvkcPg=
 github.com/libp2p/go-msgio v0.3.0 h1:mf3Z8B1xcFN314sWX+2vOTShIE0Mmn2TXn3YCUQGNj0=
@@ -413,8 +426,8 @@ github.com/libp2p/zeroconf/v2 v2.2.0 h1:Cup06Jv6u81HLhIj1KasuNM/RHHrJ8T7wOTS4+Tv
 github.com/libp2p/zeroconf/v2 v2.2.0/go.mod h1:fuJqLnUwZTshS3U/bMRJ3+ow/v9oid1n0DmyYyNO1Xs=
 github.com/lucasb-eyer/go-colorful v1.2.0 h1:1nnpGOrhyZZuNyfu1QjKiUICQ74+3FNCN69Aj6K7nkY=
 github.com/lucasb-eyer/go-colorful v1.2.0/go.mod h1:R4dSotOR9KMtayYi1e77YzuveK+i7ruzyGqttikkLy0=
-github.com/lufia/plan9stats v0.0.0-20211012122336-39d0f177ccd0 h1:6E+4a0GO5zZEnZ81pIr0yLvtUWk2if982qA3F3QD6H4=
-github.com/lufia/plan9stats v0.0.0-20211012122336-39d0f177ccd0/go.mod h1:zJYVVT2jmtg6P3p1VtQj7WsuWi/y4VnjVBn7F8KPB3I=
+github.com/lufia/plan9stats v0.0.0-20240819163618-b1d8f4d146e7 h1:5RK988zAqB3/AN3opGfRpoQgAVqr6/A5+qRTi67VUZY=
+github.com/lufia/plan9stats v0.0.0-20240819163618-b1d8f4d146e7/go.mod h1:ilwx/Dta8jXAgpFYFvSWEMwxmbWXyiUHkd5FwyKhb5k=
 github.com/lunixbochs/vtclean v1.0.0/go.mod h1:pHhQNgMf3btfWnGBVipUOjRYhoOsdGqdm/+2c2E2WMI=
 github.com/mailru/easyjson v0.0.0-20190312143242-1de009706dbe/go.mod h1:C1wdFJiN94OJF2b5HbByQZoLdCWB1Yqtg26g4irojpc=
 github.com/mailru/easyjson v0.7.7 h1:UGYAvKxe3sBsEDzO8ZeWOSlIQfWFlxbzLZe7hwFURr0=
@@ -440,8 +453,6 @@ github.com/microcosm-cc/bluemonday v1.0.26 h1:xbqSvqzQMeEHCqMi64VAs4d8uy6Mequs3r
 github.com/microcosm-cc/bluemonday v1.0.26/go.mod h1:JyzOCs9gkyQyjs+6h10UEVSe02CGwkhd72Xdqh78TWs=
 github.com/miekg/dns v1.1.41/go.mod h1:p6aan82bvRIyn+zDIv9xYNUpwa73JcSh9BKwknJysuI=
 github.com/miekg/dns v1.1.43/go.mod h1:+evo5L0630/F6ca/Z9+GAqzhjGyn8/c+TBaOyfEl0V4=
-github.com/miekg/dns v1.1.61 h1:nLxbwF3XxhwVSm8g9Dghm9MHPaUZuqhPiGL+675ZmEs=
-github.com/miekg/dns v1.1.61/go.mod h1:mnAarhS3nWaW+NVP2wTkYVIZyHNJ098SJZUki3eykwQ=
 github.com/miekg/dns v1.1.62 h1:cN8OuEF1/x5Rq6Np+h1epln8OiyPWV+lROx9LxcGgIQ=
 github.com/miekg/dns v1.1.62/go.mod h1:mvDlcItzm+br7MToIKqkglaGhlFMHJ9DTNNWONWXbNQ=
 github.com/mikioh/tcp v0.0.0-20190314235350-803a9b46060c h1:bzE/A84HN25pxAuk9Eej1Kz9OUelF97nAc82bDquQI8=
@@ -477,18 +488,18 @@ github.com/modern-go/reflect2 v1.0.1/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3Rllmb
 github.com/mr-tron/base58 v1.1.2/go.mod h1:BinMc/sQntlIE1frQmRFPUoPA1Zkr8VRgBdjWI2mNwc=
 github.com/mr-tron/base58 v1.2.0 h1:T/HDJBh4ZCPbU39/+c3rRvE0uKBQlU27+QI8LJ4t64o=
 github.com/mr-tron/base58 v1.2.0/go.mod h1:BinMc/sQntlIE1frQmRFPUoPA1Zkr8VRgBdjWI2mNwc=
-github.com/mudler/edgevpn v0.26.2 h1:OK4jfk7sYjuU7vCh+geUJk38lsxRgMk+EdsS9s0hioE=
-github.com/mudler/edgevpn v0.26.2/go.mod h1:lplntB9N6LzGNqeSM3XHCq8kyDPsNhY3jqEbWGD2WaQ=
-github.com/mudler/edgevpn v0.27.0 h1:FnBVzPs098DTgbUkiwm22n30hmEVBAq+PVpXanqx6qo=
-github.com/mudler/edgevpn v0.27.0/go.mod h1:Hwvr+i+dePgn/Yh+EMMvqcw9ByUCLAWD9TgYtJYV95Y=
-github.com/mudler/edgevpn v0.27.1 h1:UKW7/JW4l2cBAPMRnlZRHbuFDGrv7resVJlFD34WBDE=
-github.com/mudler/edgevpn v0.27.1/go.mod h1:PK7rl0QQQTdlpie9rlaS7DguH500ogqproQli/QwrxU=
-github.com/mudler/edgevpn v0.27.2 h1:FsQ95jPCDJP9LzKJYCHx70z08DGXK5yrHMzH9Qok3nE=
-github.com/mudler/edgevpn v0.27.2/go.mod h1:PK7rl0QQQTdlpie9rlaS7DguH500ogqproQli/QwrxU=
+github.com/mudler/edgevpn v0.27.4 h1:T/irkShcnU6h9OZqtvWXpNo+3gQVPUMBEoUutxJ3YUg=
+github.com/mudler/edgevpn v0.27.4/go.mod h1:NFs/RpDHCaltPFnZmOLCiUmVpTQloER1LbAtptOzqrw=
+github.com/mudler/edgevpn v0.28.0 h1:oF/Msx3zPNajy3uYLPRT5M7H3Z+sCMU0wAA8TkB11PI=
+github.com/mudler/edgevpn v0.28.0/go.mod h1:/xk8vnXUcGajPPMW5rZhPt1aD0b95LeOj2xGbRbDS8A=
+github.com/mudler/edgevpn v0.28.2 h1:wxLrH9b3NNQDgMb0Uy4gmqbGh6Ad5jdbf21GrU32xVU=
+github.com/mudler/edgevpn v0.28.2/go.mod h1:HWcdIwj5zBgOD04Hn3I+J5E5Yb3kK1CwwWaEe6/QERo=
+github.com/mudler/edgevpn v0.28.3 h1:yIuoMExwKHy/mNMBXIsm6FUFbnB9ELIxw9KXrK9KHDk=
+github.com/mudler/edgevpn v0.28.3/go.mod h1:HWcdIwj5zBgOD04Hn3I+J5E5Yb3kK1CwwWaEe6/QERo=
 github.com/mudler/go-piper v0.0.0-20240315144837-9d0100873a7d h1:8udOFrDf/I83JL0/u22j6U6Q9z9LoSdby2a/DWdd0/s=
 github.com/mudler/go-piper v0.0.0-20240315144837-9d0100873a7d/go.mod h1:O7SwdSWMilAWhBZMK9N9Y/oBDyMMzshE3ju8Xkexwig=
-github.com/mudler/go-processmanager v0.0.0-20230818213616-f204007f963c h1:CI5uGwqBpN8N7BrSKC+nmdfw+9nPQIDyjHHlaIiitZI=
-github.com/mudler/go-processmanager v0.0.0-20230818213616-f204007f963c/go.mod h1:gY3wyrhkRySJtmtI/JPt4a2mKv48h/M9pEZIW+SjeC0=
+github.com/mudler/go-processmanager v0.0.0-20240820160718-8b802d3ecf82 h1:FVT07EI8njvsD4tC2Hw8Xhactp5AWhsQWD4oTeQuSAU=
+github.com/mudler/go-processmanager v0.0.0-20240820160718-8b802d3ecf82/go.mod h1:Urp7LG5jylKoDq0663qeBh0pINGcRl35nXdKx82PSoU=
 github.com/mudler/go-stable-diffusion v0.0.0-20240429204715-4a3cd6aeae6f h1:cxtMSRkUfy+mjIQ3yMrU0txwQ4It913NEN4m1H8WWgo=
 github.com/mudler/go-stable-diffusion v0.0.0-20240429204715-4a3cd6aeae6f/go.mod h1:8ufRkpz/S/9ahkaxzZ5i4WMgO9w4InEhuRoT7vK5Rnw=
 github.com/mudler/water v0.0.0-20221010214108-8c7313014ce0 h1:Qh6ghkMgTu6siFbTf7L3IszJmshMhXxNL4V+t7IIA6w=
@@ -525,30 +536,18 @@ github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 h1:C3w9PqII01/Oq
 github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822/go.mod h1:+n7T8mK8HuQTcFwEeznm/DIxMOiR9yIdICNftLE1DvQ=
 github.com/neelance/astrewrite v0.0.0-20160511093645-99348263ae86/go.mod h1:kHJEU3ofeGjhHklVoIGuVj85JJwZ6kWPaJwCIxgnFmo=
 github.com/neelance/sourcemap v0.0.0-20151028013722-8c68805598ab/go.mod h1:Qr6/a/Q4r9LP1IltGz7tA7iOK1WonHEYhu1HRBA7ZiM=
-github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20240606155928-41c9013fa46a h1:jLmaG6BYcFvUDGFJM8B9kOM2yfvaTLxrKcFkBn4nstA=
-github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20240606155928-41c9013fa46a/go.mod h1:4T3CHXyrt+7FQHXaxULZfPjHbD8/99WuDDJa0YVZARI=
 github.com/nwaples/rardecode v1.1.0 h1:vSxaY8vQhOcVr4mm5e8XllHWTiM4JF507A0Katqw7MQ=
 github.com/nwaples/rardecode v1.1.0/go.mod h1:5DzqNKiOdpKKBH87u8VlvAnPZMXcGRhxWkRpHbbfGS0=
-github.com/nxadm/tail v1.4.4/go.mod h1:kenIhsEOeOJmVchQTgglprH7qJGnHDVpk1VPCcaMI8A=
-github.com/nxadm/tail v1.4.8/go.mod h1:+ncqLTQzXmGhMZNUePPaPqPvBxHAIsmXswZKocGu+AU=
 github.com/nxadm/tail v1.4.11 h1:8feyoE3OzPrcshW5/MJ4sGESc5cqmGkGCWlco4l0bqY=
 github.com/nxadm/tail v1.4.11/go.mod h1:OTaG3NK980DZzxbRq6lEuzgU+mug70nY11sMd4JXXHc=
 github.com/olekukonko/tablewriter v0.0.5 h1:P2Ga83D34wi1o9J6Wh1mRuqd4mF/x/lgBS7N7AbDhec=
 github.com/olekukonko/tablewriter v0.0.5/go.mod h1:hPp6KlRPjbx+hW8ykQs1w3UBbZlj6HuIJcUGPhkA7kY=
-github.com/onsi/ginkgo v1.6.0/go.mod h1:lLunBs/Ym6LB5Z9jYTR76FiuTmxDTDusOGeTQH+WWjE=
-github.com/onsi/ginkgo v1.12.1/go.mod h1:zj2OWP4+oCPe1qIXoGWkgMRwljMUYCdkwsT2108oapk=
-github.com/onsi/ginkgo v1.16.4/go.mod h1:dX+/inL/fNMqNlz0e9LfyB9TswhZpCVdJM/Z6Vvnwo0=
 github.com/onsi/ginkgo v1.16.5 h1:8xi0RTUf59SOSfEtZMvwTvXYMzG4gV23XVHOZiXNtnE=
 github.com/onsi/ginkgo v1.16.5/go.mod h1:+E8gABHa3K6zRBolWtd+ROzc/U5bkGt0FwiG042wbpU=
-github.com/onsi/ginkgo/v2 v2.19.0 h1:9Cnnf7UHo57Hy3k6/m5k3dRfGTMXGvxhHFvkDTCTpvA=
-github.com/onsi/ginkgo/v2 v2.19.0/go.mod h1:rlwLi9PilAFJ8jCg9UE1QP6VBpd6/xj3SRC0d6TU0To=
 github.com/onsi/ginkgo/v2 v2.20.0 h1:PE84V2mHqoT1sglvHc8ZdQtPcwmvvt29WLEEO3xmdZw=
 github.com/onsi/ginkgo/v2 v2.20.0/go.mod h1:lG9ey2Z29hR41WMVthyJBGUBcBhGOtoPF2VFMvBXFCI=
-github.com/onsi/gomega v1.7.1/go.mod h1:XdKZgCCFLUoM/7CFJVPcG8C1xQ1AJ0vpAezJrB7JYyY=
-github.com/onsi/gomega v1.10.1/go.mod h1:iN09h71vgCQne3DLsj+A5owkum+a2tYe+TOCB1ybHNo=
-github.com/onsi/gomega v1.16.0/go.mod h1:HnhC7FXeEQY45zxNK3PPoIUhzk/80Xly9PcubAlGdZY=
-github.com/onsi/gomega v1.33.1 h1:dsYjIxxSR755MDmKVsaFQTE22ChNBcuuTWgkUDSubOk=
-github.com/onsi/gomega v1.33.1/go.mod h1:U4R44UsT+9eLIaYRB2a5qajjtQYn0hauxvRm16AVYg0=
+github.com/onsi/ginkgo/v2 v2.20.1 h1:YlVIbqct+ZmnEph770q9Q7NVAz4wwIiVNahee6JyUzo=
+github.com/onsi/ginkgo/v2 v2.20.1/go.mod h1:lG9ey2Z29hR41WMVthyJBGUBcBhGOtoPF2VFMvBXFCI=
 github.com/onsi/gomega v1.34.1 h1:EUMJIKUjM8sKjYbtxQI9A4z2o+rruxnzNvpknOXie6k=
 github.com/onsi/gomega v1.34.1/go.mod h1:kU1QgUvBDLXBJq618Xvm2LUX6rSAfRaFRTcdOeDLwwY=
 github.com/opencontainers/go-digest v1.0.0 h1:apOUWs51W5PlhuyGyz9FCeeBIOUDA/6nW8Oi/yOhh5U=
@@ -579,21 +578,13 @@ github.com/philhofer/fwd v1.1.2 h1:bnDivRJ1EWPjUIRXV5KfORO897HTbpFAQddBdE8t7Gw=
 github.com/philhofer/fwd v1.1.2/go.mod h1:qkPdfjR2SIEbspLqpe1tO4n5yICnr2DY7mqEx2tUTP0=
 github.com/pierrec/lz4/v4 v4.1.2 h1:qvY3YFXRQE/XB8MlLzJH7mSzBs74eA2gg52YTk6jUPM=
 github.com/pierrec/lz4/v4 v4.1.2/go.mod h1:gZWDp/Ze/IJXGXf23ltt2EXimqmTUXEy0GFuRQyBid4=
-github.com/pion/datachannel v1.5.6 h1:1IxKJntfSlYkpUj8LlYRSWpYiTTC02nUrOE8T3DqGeg=
-github.com/pion/datachannel v1.5.6/go.mod h1:1eKT6Q85pRnr2mHiWHxJwO50SfZRtWHTsNIVb/NfGW4=
 github.com/pion/datachannel v1.5.8 h1:ph1P1NsGkazkjrvyMfhRBUAWMxugJjq2HfQifaOoSNo=
 github.com/pion/datachannel v1.5.8/go.mod h1:PgmdpoaNBLX9HNzNClmdki4DYW5JtI7Yibu8QzbL3tI=
 github.com/pion/dtls/v2 v2.2.7/go.mod h1:8WiMkebSHFD0T+dIU+UeBaoV7kDhOW5oDCzZ7WZ/F9s=
-github.com/pion/dtls/v2 v2.2.11 h1:9U/dpCYl1ySttROPWJgqWKEylUdT0fXp/xst6JwY5Ks=
-github.com/pion/dtls/v2 v2.2.11/go.mod h1:d9SYc9fch0CqK90mRk1dC7AkzzpwJj6u2GU3u+9pqFE=
 github.com/pion/dtls/v2 v2.2.12 h1:KP7H5/c1EiVAAKUmXyCzPiQe5+bCJrpOeKg/L05dunk=
 github.com/pion/dtls/v2 v2.2.12/go.mod h1:d9SYc9fch0CqK90mRk1dC7AkzzpwJj6u2GU3u+9pqFE=
-github.com/pion/ice/v2 v2.3.25 h1:M5rJA07dqhi3nobJIg+uPtcVjFECTrhcR3n0ns8kDZs=
-github.com/pion/ice/v2 v2.3.25/go.mod h1:KXJJcZK7E8WzrBEYnV4UtqEZsGeWfHxsNqhVcVvgjxw=
 github.com/pion/ice/v2 v2.3.34 h1:Ic1ppYCj4tUOcPAp76U6F3fVrlSw8A9JtRXLqw6BbUM=
 github.com/pion/ice/v2 v2.3.34/go.mod h1:mBF7lnigdqgtB+YHkaY/Y6s6tsyRyo4u4rPGRuOjUBQ=
-github.com/pion/interceptor v0.1.29 h1:39fsnlP1U8gw2JzOFWdfCU82vHvhW9o0rZnZF56wF+M=
-github.com/pion/interceptor v0.1.29/go.mod h1:ri+LGNjRUc5xUNtDEPzfdkmSqISixVTBF/z/Zms/6T4=
 github.com/pion/interceptor v0.1.30 h1:au5rlVHsgmxNi+v/mjOPazbW1SHzfx7/hYOEYQnUcxA=
 github.com/pion/interceptor v0.1.30/go.mod h1:RQuKT5HTdkP2Fi0cuOS5G5WNymTjzXaGF75J4k7z2nc=
 github.com/pion/logging v0.2.2 h1:M9+AIj/+pxNsDfAT64+MAVgJO0rsyLnoJKCqf//DoeY=
@@ -606,39 +597,27 @@ github.com/pion/rtcp v1.2.12/go.mod h1:sn6qjxvnwyAkkPzPULIbVqSKI5Dv54Rv7VG0kNxh9
 github.com/pion/rtcp v1.2.14 h1:KCkGV3vJ+4DAJmvP0vaQShsb0xkRfWkO540Gy102KyE=
 github.com/pion/rtcp v1.2.14/go.mod h1:sn6qjxvnwyAkkPzPULIbVqSKI5Dv54Rv7VG0kNxh9L4=
 github.com/pion/rtp v1.8.3/go.mod h1:pBGHaFt/yW7bf1jjWAoUjpSNoDnw98KTMg+jWWvziqU=
-github.com/pion/rtp v1.8.6 h1:MTmn/b0aWWsAzux2AmP8WGllusBVw4NPYPVFFd7jUPw=
-github.com/pion/rtp v1.8.6/go.mod h1:pBGHaFt/yW7bf1jjWAoUjpSNoDnw98KTMg+jWWvziqU=
 github.com/pion/rtp v1.8.9 h1:E2HX740TZKaqdcPmf4pw6ZZuG8u5RlMMt+l3dxeu6Wk=
 github.com/pion/rtp v1.8.9/go.mod h1:pBGHaFt/yW7bf1jjWAoUjpSNoDnw98KTMg+jWWvziqU=
-github.com/pion/sctp v1.8.13/go.mod h1:YKSgO/bO/6aOMP9LCie1DuD7m+GamiK2yIiPM6vH+GA=
-github.com/pion/sctp v1.8.16 h1:PKrMs+o9EMLRvFfXq59WFsC+V8mN1wnKzqrv+3D/gYY=
-github.com/pion/sctp v1.8.16/go.mod h1:P6PbDVA++OJMrVNg2AL3XtYHV4uD6dvfyOovCgMs0PE=
 github.com/pion/sctp v1.8.33 h1:dSE4wX6uTJBcNm8+YlMg7lw1wqyKHggsP5uKbdj+NZw=
 github.com/pion/sctp v1.8.33/go.mod h1:beTnqSzewI53KWoG3nqB282oDMGrhNxBdb+JZnkCwRM=
 github.com/pion/sdp/v3 v3.0.9 h1:pX++dCHoHUwq43kuwf3PyJfHlwIj4hXA7Vrifiq0IJY=
 github.com/pion/sdp/v3 v3.0.9/go.mod h1:B5xmvENq5IXJimIO4zfp6LAe1fD9N+kFv+V/1lOdz8M=
-github.com/pion/srtp/v2 v2.0.18 h1:vKpAXfawO9RtTRKZJbG4y0v1b11NZxQnxRl85kGuUlo=
-github.com/pion/srtp/v2 v2.0.18/go.mod h1:0KJQjA99A6/a0DOVTu1PhDSw0CXF2jTkqOoMg3ODqdA=
 github.com/pion/srtp/v2 v2.0.20 h1:HNNny4s+OUmG280ETrCdgFndp4ufx3/uy85EawYEhTk=
 github.com/pion/srtp/v2 v2.0.20/go.mod h1:0KJQjA99A6/a0DOVTu1PhDSw0CXF2jTkqOoMg3ODqdA=
 github.com/pion/stun v0.6.1 h1:8lp6YejULeHBF8NmV8e2787BogQhduZugh5PdhDyyN4=
 github.com/pion/stun v0.6.1/go.mod h1:/hO7APkX4hZKu/D0f2lHzNyvdkTGtIy3NDmLR7kSz/8=
 github.com/pion/transport/v2 v2.2.1/go.mod h1:cXXWavvCnFF6McHTft3DWS9iic2Mftcz1Aq29pGcU5g=
-github.com/pion/transport/v2 v2.2.2/go.mod h1:OJg3ojoBJopjEeECq2yJdXH9YVrUJ1uQ++NjXLOUorc=
 github.com/pion/transport/v2 v2.2.3/go.mod h1:q2U/tf9FEfnSBGSW6w5Qp5PFWRLRj3NjLhCCgpRK4p0=
 github.com/pion/transport/v2 v2.2.4/go.mod h1:q2U/tf9FEfnSBGSW6w5Qp5PFWRLRj3NjLhCCgpRK4p0=
-github.com/pion/transport/v2 v2.2.5 h1:iyi25i/21gQck4hfRhomF6SktmUQjRsRW4WJdhfc3Kc=
-github.com/pion/transport/v2 v2.2.5/go.mod h1:q2U/tf9FEfnSBGSW6w5Qp5PFWRLRj3NjLhCCgpRK4p0=
 github.com/pion/transport/v2 v2.2.10 h1:ucLBLE8nuxiHfvkFKnkDQRYWYfp8ejf4YBOPfaQpw6Q=
 github.com/pion/transport/v2 v2.2.10/go.mod h1:sq1kSLWs+cHW9E+2fJP95QudkzbK7wscs8yYgQToO5E=
 github.com/pion/transport/v3 v3.0.1/go.mod h1:UY7kiITrlMv7/IKgd5eTUcaahZx5oUN3l9SzK5f5xE0=
-github.com/pion/transport/v3 v3.0.2 h1:r+40RJR25S9w3jbA6/5uEPTzcdn7ncyU44RWCbHkLg4=
-github.com/pion/transport/v3 v3.0.2/go.mod h1:nIToODoOlb5If2jF9y2Igfx3PFYWfuXi37m0IlWa/D0=
+github.com/pion/transport/v3 v3.0.7 h1:iRbMH05BzSNwhILHoBoAPxoB9xQgOaJk+591KC9P1o0=
+github.com/pion/transport/v3 v3.0.7/go.mod h1:YleKiTZ4vqNxVwh77Z0zytYi7rXHl7j6uPLGhhz9rwo=
 github.com/pion/turn/v2 v2.1.3/go.mod h1:huEpByKKHix2/b9kmTAM3YoX6MKP+/D//0ClgUYR2fY=
 github.com/pion/turn/v2 v2.1.6 h1:Xr2niVsiPTB0FPtt+yAWKFUkU1eotQbGgpTIld4x1Gc=
 github.com/pion/turn/v2 v2.1.6/go.mod h1:huEpByKKHix2/b9kmTAM3YoX6MKP+/D//0ClgUYR2fY=
-github.com/pion/webrtc/v3 v3.2.40 h1:Wtfi6AZMQg+624cvCXUuSmrKWepSB7zfgYDOYqsSOVU=
-github.com/pion/webrtc/v3 v3.2.40/go.mod h1:M1RAe3TNTD1tzyvqHrbVODfwdPGSXOUo/OgpoGGJqFY=
 github.com/pion/webrtc/v3 v3.3.0 h1:Rf4u6n6U5t5sUxhYPQk/samzU/oDv7jk6BA5hyO2F9I=
 github.com/pion/webrtc/v3 v3.3.0/go.mod h1:hVmrDJvwhEertRWObeb1xzulzHGeVUoPlWvxdGzcfU0=
 github.com/pkg/errors v0.8.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
@@ -650,11 +629,9 @@ github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZb
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/polydawn/refmt v0.89.0 h1:ADJTApkvkeBZsN0tBTx8QjpD9JkmxbKp0cxfr9qszm4=
 github.com/polydawn/refmt v0.89.0/go.mod h1:/zvteZs/GwLtCgZ4BL6CBsk9IKIlexP43ObX9AxTqTw=
-github.com/power-devops/perfstat v0.0.0-20210106213030-5aafc221ea8c h1:ncq/mPwQF4JjgDlrVEn3C11VoGHZN7m8qihwgMEtzYw=
-github.com/power-devops/perfstat v0.0.0-20210106213030-5aafc221ea8c/go.mod h1:OmDBASR4679mdNQnz2pUhc2G8CO2JrUAVFDRBDP/hJE=
+github.com/power-devops/perfstat v0.0.0-20240221224432-82ca36839d55 h1:o4JXh1EVt9k/+g42oCprj/FisM4qX9L3sZB3upGN2ZU=
+github.com/power-devops/perfstat v0.0.0-20240221224432-82ca36839d55/go.mod h1:OmDBASR4679mdNQnz2pUhc2G8CO2JrUAVFDRBDP/hJE=
 github.com/prometheus/client_golang v0.8.0/go.mod h1:7SWBe2y4D6OKWSNQJUaRYU/AaXPKyh/dDVn+NZz0KFw=
-github.com/prometheus/client_golang v1.19.1 h1:wZWJDwK+NameRJuPGDhlnFgx8e8HN3XHQeLaYJFJBOE=
-github.com/prometheus/client_golang v1.19.1/go.mod h1:mP78NwGzrVks5S2H6ab8+ZZGJLZUq1hoULYBAYBw1Ho=
 github.com/prometheus/client_golang v1.20.0 h1:jBzTZ7B099Rg24tny+qngoynol8LtVYlA2bqx3vEloI=
 github.com/prometheus/client_golang v1.20.0/go.mod h1:PIEt8X02hGcP8JWbeHyeZ53Y/jReSnHgO035n//V5WE=
 github.com/prometheus/client_model v0.0.0-20180712105110-5c3871d89910/go.mod h1:MbSGuTsp3dbXC40dX6PRTWyKYBIrTGTE9sqQNg2J8bo=
@@ -669,8 +646,6 @@ github.com/prometheus/procfs v0.15.1 h1:YagwOFzUgYfKKHX6Dr+sHT7km/hxC76UB0leargg
 github.com/prometheus/procfs v0.15.1/go.mod h1:fB45yRUv8NstnjriLhBQLuOUt+WW4BsoGhij/e3PBqk=
 github.com/quic-go/qpack v0.4.0 h1:Cr9BXA1sQS2SmDUWjSofMPNKmvF6IiIfDRmgU0w1ZCo=
 github.com/quic-go/qpack v0.4.0/go.mod h1:UZVnYIfi5GRk+zI9UMaCPsmZ2xKJP7XBUvVyT1Knj9A=
-github.com/quic-go/quic-go v0.44.0 h1:So5wOr7jyO4vzL2sd8/pD9Kesciv91zSk8BoFngItQ0=
-github.com/quic-go/quic-go v0.44.0/go.mod h1:z4cx/9Ny9UtGITIPzmPTXh1ULfOyWh4qGQlpnPcWmek=
 github.com/quic-go/quic-go v0.46.0 h1:uuwLClEEyk1DNvchH8uCByQVjo3yKL9opKulExNDs7Y=
 github.com/quic-go/quic-go v0.46.0/go.mod h1:1dLehS7TIR64+vxGR70GDcatWTOtMX2PUtnKsjbTurI=
 github.com/quic-go/webtransport-go v0.8.0 h1:HxSrwun11U+LlmwpgM1kEqIqH90IT4N8auv/cD7QFJg=
@@ -698,9 +673,10 @@ github.com/sashabaranov/go-openai v1.26.2/go.mod h1:lj5b/K+zjTSFxVLijLSTDZuP7adO
 github.com/schollz/progressbar/v3 v3.14.4 h1:W9ZrDSJk7eqmQhd3uxFNNcTr0QL+xuGNI9dEMrw0r74=
 github.com/schollz/progressbar/v3 v3.14.4/go.mod h1:aT3UQ7yGm+2ZjeXPqsjTenwL3ddUiuZ0kfQ/2tHlyNI=
 github.com/sergi/go-diff v1.0.0/go.mod h1:0CfEIISq7TuYL3j771MWULgwwjU+GofnZX9QAmXWZgo=
-github.com/shirou/gopsutil/v3 v3.23.7/go.mod h1:c4gnmoRC0hQuaLqvxnx1//VXQ0Ms/X9UnJF8pddY5z4=
 github.com/shirou/gopsutil/v3 v3.24.5 h1:i0t8kL+kQTvpAYToeuiVk3TgDeKOFioZO3Ztz/iZ9pI=
 github.com/shirou/gopsutil/v3 v3.24.5/go.mod h1:bsoOS1aStSs9ErQ1WWfxllSeS1K5D+U30r2NfcubMVk=
+github.com/shirou/gopsutil/v4 v4.24.7 h1:V9UGTK4gQ8HvcnPKf6Zt3XHyQq/peaekfxpJ2HSocJk=
+github.com/shirou/gopsutil/v4 v4.24.7/go.mod h1:0uW/073rP7FYLOkvxolUQM5rMOLTNmRXnFKafpb71rw=
 github.com/shoenig/go-m1cpu v0.1.6 h1:nxdKQNcEB6vzgA2E2bvzKIYRuNj7XNJ4S/aRSwKzFtM=
 github.com/shoenig/go-m1cpu v0.1.6/go.mod h1:1JJMcUBvfNwpq05QDQVAnx3gUHr9IYF7GNg9SUEw2VQ=
 github.com/shoenig/test v0.6.4 h1:kVTaSd7WLz5WZ2IaoM0RSzRsUD+m8wRR+5qvntpn4LU=
@@ -776,12 +752,10 @@ github.com/thxcode/gguf-parser-go v0.1.0 h1:J4QruXyEQGjrAKeKZFlsD2na9l4XF5+bjR19
 github.com/thxcode/gguf-parser-go v0.1.0/go.mod h1:Tn1PsO/YDEtLIxm1+QDCjIIH9L/9Sr7+KpxZKm0sEuE=
 github.com/tinylib/msgp v1.1.8 h1:FCXC1xanKO4I8plpHGH2P7koL/RzZs12l/+r7vakfm0=
 github.com/tinylib/msgp v1.1.8/go.mod h1:qkpG+2ldGg4xRFmx+jfTvZPxfGFhi64BcnL9vkCm/Tw=
-github.com/tklauser/go-sysconf v0.3.11/go.mod h1:GqXfhXY3kiPa0nAXPDIQIWzJbMCB7AmcWpGR8lSZfqI=
-github.com/tklauser/go-sysconf v0.3.12 h1:0QaGUFOdQaIVdPgfITYzaTegZvdCjmYO52cSFAEVmqU=
-github.com/tklauser/go-sysconf v0.3.12/go.mod h1:Ho14jnntGE1fpdOqQEEaiKRpvIavV0hSfmBq8nJbHYI=
-github.com/tklauser/numcpus v0.6.0/go.mod h1:FEZLMke0lhOUG6w2JadTzp0a+Nl8PF/GFkQ5UVIcaL4=
-github.com/tklauser/numcpus v0.6.1 h1:ng9scYS7az0Bk4OZLvrNXNSAO2Pxr1XXRAPyjhIx+Fk=
-github.com/tklauser/numcpus v0.6.1/go.mod h1:1XfjsgE2zo8GVw7POkMbHENHzVg3GzmoZ9fESEdAacY=
+github.com/tklauser/go-sysconf v0.3.14 h1:g5vzr9iPFFz24v2KZXs/pvpvh8/V9Fw6vQK5ZZb78yU=
+github.com/tklauser/go-sysconf v0.3.14/go.mod h1:1ym4lWMLUOhuBOPGtRcJm7tEGX4SCYNEEEtghGG/8uY=
+github.com/tklauser/numcpus v0.8.0 h1:Mx4Wwe/FjZLeQsK/6kt2EOepwwSl7SmJrK5bV/dXYgY=
+github.com/tklauser/numcpus v0.8.0/go.mod h1:ZJZlAY+dmR4eut8epnzf0u/VwodKmryxR8txiloSqBE=
 github.com/tmc/langchaingo v0.1.12 h1:yXwSu54f3b1IKw0jJ5/DWu+qFVH1NBblwC0xddBzGJE=
 github.com/tmc/langchaingo v0.1.12/go.mod h1:cd62xD6h+ouk8k/QQFhOsjRYBSA1JJ5UVKXSIgm7Ni4=
 github.com/ulikunitz/xz v0.5.8/go.mod h1:nbz6k7qbPmH4IRqmfOplQw/tblSgqTqBwxkY0oWt/14=
@@ -791,8 +765,6 @@ github.com/urfave/cli v1.22.2/go.mod h1:Gos4lmkARVdJ6EkW0WaNv/tZAAMe9V7XWyB60NtX
 github.com/urfave/cli v1.22.10/go.mod h1:Gos4lmkARVdJ6EkW0WaNv/tZAAMe9V7XWyB60NtXRu0=
 github.com/urfave/cli v1.22.12 h1:igJgVw1JdKH+trcLWLeLwZjU9fEfPesQ+9/e4MQ44S8=
 github.com/urfave/cli v1.22.12/go.mod h1:sSBEIC79qR6OvcmsD4U3KABeOTxDqQtdDnaFuUN30b8=
-github.com/urfave/cli/v2 v2.27.3 h1:/POWahRmdh7uztQ3CYnaDddk0Rm90PyOgIxgW2rr41M=
-github.com/urfave/cli/v2 v2.27.3/go.mod h1:m4QzxcD2qpra4z7WhzEGn74WZLViBnMpb1ToCAKdGRQ=
 github.com/urfave/cli/v2 v2.27.4 h1:o1owoI+02Eb+K107p27wEX9Bb8eqIoZCfLXloLUSWJ8=
 github.com/urfave/cli/v2 v2.27.4/go.mod h1:m4QzxcD2qpra4z7WhzEGn74WZLViBnMpb1ToCAKdGRQ=
 github.com/valyala/bytebufferpool v1.0.0 h1:GqA5TC/0021Y/b9FG4Oi9Mr3q7XYx6KllzawFIhcdPw=
@@ -809,9 +781,13 @@ github.com/viant/assertly v0.4.8/go.mod h1:aGifi++jvCrUaklKEKT0BU95igDNaqkvz+49u
 github.com/viant/toolbox v0.24.0/go.mod h1:OxMCG57V0PXuIP2HNQrtJf2CjqdmbrOx5EkMILuUhzM=
 github.com/vishvananda/netlink v1.2.1-beta.2 h1:Llsql0lnQEbHj0I1OuKyp8otXp0r3q0mPkuhwHfStVs=
 github.com/vishvananda/netlink v1.2.1-beta.2/go.mod h1:twkDnbuQxJYemMlGd4JFIcuhgX83tXhKS2B/PRMpOho=
+github.com/vishvananda/netlink v1.3.0 h1:X7l42GfcV4S6E4vHTsw48qbrV+9PVojNfIhZcwQdrZk=
+github.com/vishvananda/netlink v1.3.0/go.mod h1:i6NetklAujEcC6fK0JPjT8qSwWyO0HLn4UKG+hGqeJs=
 github.com/vishvananda/netns v0.0.0-20200728191858-db3c7e526aae/go.mod h1:DD4vA1DwXk04H54A1oHXtwZmA0grkVMdPxx/VGLCah0=
 github.com/vishvananda/netns v0.0.0-20211101163701-50045581ed74 h1:gga7acRE695APm9hlsSMoOoE65U4/TcqNj90mc69Rlg=
 github.com/vishvananda/netns v0.0.0-20211101163701-50045581ed74/go.mod h1:DD4vA1DwXk04H54A1oHXtwZmA0grkVMdPxx/VGLCah0=
+github.com/vishvananda/netns v0.0.4 h1:Oeaw1EM2JMxD51g9uhtC0D7erkIjgmj8+JZc26m1YX8=
+github.com/vishvananda/netns v0.0.4/go.mod h1:SpkAiCQRtJ6TvvxPnOSyH3BMl6unz3xZlaprSwhNNJM=
 github.com/warpfork/go-wish v0.0.0-20220906213052-39a1cc7a02d0 h1:GDDkbFiaK8jsSDJfjId/PEGEShv6ugrt4kYsC5UIDaQ=
 github.com/warpfork/go-wish v0.0.0-20220906213052-39a1cc7a02d0/go.mod h1:x6AKhvSSexNrVSrViXSHUEbICjmGXhtgABaHIySUSGw=
 github.com/whyrusleeping/go-keyspace v0.0.0-20160322163242-5b898ac5add1 h1:EKhdznlJHPMoKr0XTrX+IlJs1LH3lyx2nfr1dOlZ79k=
@@ -839,12 +815,13 @@ github.com/yuin/goldmark v1.5.4 h1:2uY/xC0roWy8IBEGLgB1ywIoEJFGmRrX21YQcvGZzjU=
 github.com/yuin/goldmark v1.5.4/go.mod h1:6yULJ656Px+3vBD8DxQVa3kxgyrAnzto9xy5taEt/CY=
 github.com/yuin/goldmark-emoji v1.0.2 h1:c/RgTShNgHTtc6xdz2KKI74jJr6rWi7FPgnP9GAsO5s=
 github.com/yuin/goldmark-emoji v1.0.2/go.mod h1:RhP/RWpexdp+KHs7ghKnifRoIs/Bq4nDS7tRbCkOwKY=
-github.com/yusufpapurcu/wmi v1.2.3/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0=
 github.com/yusufpapurcu/wmi v1.2.4 h1:zFUKzehAFReQwLys1b/iSMl+JQGSCSjtVqQn9bBrPo0=
 github.com/yusufpapurcu/wmi v1.2.4/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0=
 go.opencensus.io v0.18.0/go.mod h1:vKdFvxhtzZ9onBp9VKHK8z/sRpBMnKAsufL7wlDrCOA=
 go.opencensus.io v0.24.0 h1:y73uSU6J157QMP2kn2r30vwW1A2W2WFwSCGnAVxeaD0=
 go.opencensus.io v0.24.0/go.mod h1:vNK8G9p7aAivkbmorf4v+7Hgx+Zs0yY+0fOtgBfjQKo=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.52.0 h1:9l89oX4ba9kHbBol3Xin3leYJ+252h0zszDtBwyKe2A=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.52.0/go.mod h1:XLZfZboOJWHNKUv7eH0inh0E9VV6eWDFB/9yJyTLPp0=
 go.opentelemetry.io/otel v1.28.0 h1:/SqNcYk+idO0CxKEUOtKQClMK/MimZihKYMruSMViUo=
 go.opentelemetry.io/otel v1.28.0/go.mod h1:q68ijF8Fc8CnMHKyzqL6akLO46ePnjkgfIMIjUIX9z4=
 go.opentelemetry.io/otel/exporters/prometheus v0.50.0 h1:2Ewsda6hejmbhGFyUvWZjUThC98Cf8Zy6g0zkIimOng=
@@ -859,12 +836,8 @@ go.opentelemetry.io/otel/trace v1.28.0 h1:GhQ9cUuQGmNDd5BTCP2dAvv75RdMxEfTmYejp+
 go.opentelemetry.io/otel/trace v1.28.0/go.mod h1:jPyXzNPg6da9+38HEwElrQiHlVMTnVfM3/yv2OlIHaI=
 go.uber.org/atomic v1.6.0/go.mod h1:sABNBOSYdrvTF6hTgEIbc7YasKWGhgEQZyfxyTvoXHQ=
 go.uber.org/atomic v1.7.0/go.mod h1:fEN4uk6kAWBTFdckzkM89CLk9XfWZrxpCo0nPH17wJc=
-go.uber.org/dig v1.17.1 h1:Tga8Lz8PcYNsWsyHMZ1Vm0OQOUaJNDyvPImgbAu9YSc=
-go.uber.org/dig v1.17.1/go.mod h1:Us0rSJiThwCv2GteUN0Q7OKvU7n5J4dxZ9JKUXozFdE=
 go.uber.org/dig v1.18.0 h1:imUL1UiY0Mg4bqbFfsRQO5G4CGRBec/ZujWTvSVp3pw=
 go.uber.org/dig v1.18.0/go.mod h1:Us0rSJiThwCv2GteUN0Q7OKvU7n5J4dxZ9JKUXozFdE=
-go.uber.org/fx v1.22.1 h1:nvvln7mwyT5s1q201YE29V/BFrGor6vMiDNpU/78Mys=
-go.uber.org/fx v1.22.1/go.mod h1:HT2M7d7RHo+ebKGh9NRcrsrHHfpZ60nW3QRubMRfv48=
 go.uber.org/fx v1.22.2 h1:iPW+OPxv0G8w75OemJ1RAnTUrF55zOJlXlo1TbJ0Buw=
 go.uber.org/fx v1.22.2/go.mod h1:o/D9n+2mLP6v1EG+qsdT1O8wKopYAsqZasju97SDFCU=
 go.uber.org/goleak v1.1.11-0.20210813005559-691160354723/go.mod h1:cwTWslyiVhfpKIDGSZEM2HlOvcqm+tG4zioyIeLoqMQ=
@@ -895,20 +868,11 @@ golang.org/x/crypto v0.0.0-20210322153248-0c34fe9e7dc2/go.mod h1:T9bdIzuCu7OtxOm
 golang.org/x/crypto v0.0.0-20210921155107-089bfa567519/go.mod h1:GvvjBRRGRdwPK5ydBHafDWAxML/pGHZbMvKqRZ5+Abc=
 golang.org/x/crypto v0.3.0/go.mod h1:hebNnKkNXi2UzZN1eVRvBB7co0a+JxK6XbPiWVs/3J4=
 golang.org/x/crypto v0.8.0/go.mod h1:mRqEX+O9/h5TFCrQhkgjo2yKi0yYA+9ecGkdQoHrywE=
-golang.org/x/crypto v0.11.0/go.mod h1:xgJhtzW8F9jGdVFWZESrid1U1bjeNy4zgy5cRr/CIio=
 golang.org/x/crypto v0.12.0/go.mod h1:NF0Gs7EO5K4qLn+Ylc+fih8BSTeIjAP05siRnAh98yw=
 golang.org/x/crypto v0.18.0/go.mod h1:R0j02AL6hcrfOiy9T4ZYp/rcWeMxM3L6QYxlOuEG1mg=
-golang.org/x/crypto v0.19.0/go.mod h1:Iy9bg/ha4yyC70EfRS8jz+B6ybOBKMaSxLj6P6oBDfU=
-golang.org/x/crypto v0.21.0/go.mod h1:0BP7YvVV9gBbVKyeTG0Gyn+gZm94bibOW5BjDEYAOMs=
-golang.org/x/crypto v0.24.0 h1:mnl8DM0o513X8fdIkmyFE/5hTYxbwYOjDS/+rK6qpRI=
-golang.org/x/crypto v0.24.0/go.mod h1:Z1PMYSOR5nyMcyAVAIQSKCDwalqy85Aqn1x3Ws4L5DM=
 golang.org/x/crypto v0.26.0 h1:RrRspgV4mU+YwB4FYnuBoKsUapNIL5cohGAmSH3azsw=
 golang.org/x/crypto v0.26.0/go.mod h1:GY7jblb9wI+FOo5y8/S2oY4zWP07AkOJ4+jxCqdqn54=
 golang.org/x/exp v0.0.0-20190121172915-509febef88a4/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
-golang.org/x/exp v0.0.0-20240506185415-9bf2ced13842 h1:vr/HnozRka3pE4EsMEg1lgkXJkTFJCVUX+S/ZT6wYzM=
-golang.org/x/exp v0.0.0-20240506185415-9bf2ced13842/go.mod h1:XtvwrStGgqGPLc4cjQfWqZHG1YFdYs6swckp8vpsjnc=
-golang.org/x/exp v0.0.0-20240719175910-8a7402abbf56 h1:2dVuKD2vS7b0QIHQbpyTISPd0LeHDbnYEryqj5Q1ug8=
-golang.org/x/exp v0.0.0-20240719175910-8a7402abbf56/go.mod h1:M4RDyNAINzryxdtnbRXRL/OHtkFuWGRjvuhBJpk2IlY=
 golang.org/x/exp v0.0.0-20240808152545-0cdaa3abc0fa h1:ELnwvuAXPNtPk1TJRuGkI9fDTwym6AYBu0qzT8AcHdI=
 golang.org/x/exp v0.0.0-20240808152545-0cdaa3abc0fa/go.mod h1:akd2r19cwCdwSwWeIdzYQGa/EZZyqcOdwWiwj5L5eKQ=
 golang.org/x/lint v0.0.0-20180702182130-06c8688daad7/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
@@ -925,8 +889,6 @@ golang.org/x/mod v0.4.2/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
 golang.org/x/mod v0.6.0-dev.0.20220419223038-86c51ed26bb4/go.mod h1:jJ57K6gSWd91VN4djpZkiMVwK6gcyfeH4XE8wZrZaV4=
 golang.org/x/mod v0.7.0/go.mod h1:iBbtSCu2XBx23ZKBPSOrRkjjQPZFPuis4dIYUhu/chs=
 golang.org/x/mod v0.8.0/go.mod h1:iBbtSCu2XBx23ZKBPSOrRkjjQPZFPuis4dIYUhu/chs=
-golang.org/x/mod v0.18.0 h1:5+9lSbEzPSdWkH32vYPBwEpX8KwDbM52Ud9xBUvNlb0=
-golang.org/x/mod v0.18.0/go.mod h1:hTbmBsO62+eylJbnUtE2MGJUyE7QWk4xUqPFrRgJ+7c=
 golang.org/x/mod v0.20.0 h1:utOm6MM3R3dnawAiJgn0y+xvuYRsm1RKM/4giyfDgV0=
 golang.org/x/mod v0.20.0/go.mod h1:hTbmBsO62+eylJbnUtE2MGJUyE7QWk4xUqPFrRgJ+7c=
 golang.org/x/net v0.0.0-20180724234803-3673e40ba225/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
@@ -941,33 +903,28 @@ golang.org/x/net v0.0.0-20190313220215-9f648a60d977/go.mod h1:t9HGtf8HONx5eT2rtn
 golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
 golang.org/x/net v0.0.0-20190620200207-3b0461eec859/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
 golang.org/x/net v0.0.0-20200226121028-0de0cce0169b/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
-golang.org/x/net v0.0.0-20200520004742-59133d7f0dd7/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
 golang.org/x/net v0.0.0-20201021035429-f5854403a974/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
 golang.org/x/net v0.0.0-20201110031124-69a78807bb2b/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
 golang.org/x/net v0.0.0-20210119194325-5f4716e94777/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
 golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
 golang.org/x/net v0.0.0-20210405180319-a5a99cb37ef4/go.mod h1:p54w0d4576C0XHj96bSt6lcn1PtDYWL6XObtHCRCNQM=
 golang.org/x/net v0.0.0-20210423184538-5f58ad60dda6/go.mod h1:OJAsFXCWl8Ukc7SiCT/9KSuxbyM7479/AVlXFRxuMCk=
-golang.org/x/net v0.0.0-20210428140749-89ef3d95e781/go.mod h1:OJAsFXCWl8Ukc7SiCT/9KSuxbyM7479/AVlXFRxuMCk=
 golang.org/x/net v0.0.0-20220722155237-a158d28d115b/go.mod h1:XRhObCWvk6IyKnWLug+ECip1KBveYUHfp+8e9klMJ9c=
 golang.org/x/net v0.2.0/go.mod h1:KqCZLdyyvdV855qA2rE3GC2aiw5xGR5TEjj8smXukLY=
 golang.org/x/net v0.3.0/go.mod h1:MBQ8lrhLObU/6UmLb4fmbmk5OcyYmqtbGd/9yIeKjEE=
 golang.org/x/net v0.6.0/go.mod h1:2Tu9+aMcznHK/AK1HMvgo6xiTLG5rD5rZLDS+rp2Bjs=
 golang.org/x/net v0.9.0/go.mod h1:d48xBJpPfHeWQsugry2m+kC02ZBRGRgulfHnEXEuWns=
 golang.org/x/net v0.10.0/go.mod h1:0qNGK6F8kojg2nk9dLZ2mShWaEBan6FAoqfSigmmuDg=
-golang.org/x/net v0.13.0/go.mod h1:zEVYFnQC7m/vmpQFELhcD1EWkZlX69l4oqgmer6hfKA=
 golang.org/x/net v0.14.0/go.mod h1:PpSgVXXLK0OxS0F31C1/tv6XNguvCrnXIDrFMspZIUI=
 golang.org/x/net v0.20.0/go.mod h1:z8BVo6PvndSri0LbOE3hAn0apkU+1YvI6E70E9jsnvY=
-golang.org/x/net v0.21.0/go.mod h1:bIjVDfnllIU7BJ2DNgfnXvpSvtn8VRwhlsaeUTyUS44=
-golang.org/x/net v0.22.0/go.mod h1:JKghWKKOSdJwpW2GEx0Ja7fmaKnMsbu+MWVZTokSYmg=
-golang.org/x/net v0.26.0 h1:soB7SVo0PWrY4vPW/+ay0jKDNScG2X9wFeYlXIvJsOQ=
-golang.org/x/net v0.26.0/go.mod h1:5YKkiSynbBIh3p6iOc/vibscux0x38BZDkn8sCUPxHE=
 golang.org/x/net v0.28.0 h1:a9JDOJc5GMUJ0+UDqmLT86WiEy7iWyIhz8gz8E4e5hE=
 golang.org/x/net v0.28.0/go.mod h1:yqtgsTWOOnlGLG9GFRrK3++bGOUEkNBoHZc8MEDWPNg=
 golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
 golang.org/x/oauth2 v0.0.0-20181017192945-9dcd33a902f4/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
 golang.org/x/oauth2 v0.0.0-20181203162652-d668ce993890/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
 golang.org/x/oauth2 v0.0.0-20190226205417-e64efc72b421/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
+golang.org/x/oauth2 v0.21.0 h1:tsimM75w1tF/uws5rbeHzIWxEqElMehnc+iW793zsZs=
+golang.org/x/oauth2 v0.21.0/go.mod h1:XYTD2NtWslqkgxebSiOHnXEap4TF09sJSc7H1sXbhtI=
 golang.org/x/perf v0.0.0-20180704124530-6e6d33e29852/go.mod h1:JLpeXjPJfIyPr5TlbXLkXWLhP8nz10XfvxElABhCtcw=
 golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20181108010431-42b317875d0f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
@@ -979,8 +936,6 @@ golang.org/x/sync v0.0.0-20201020160332-67f06af15bc9/go.mod h1:RxMgew5VJxzue5/jJ
 golang.org/x/sync v0.0.0-20210220032951-036812b2e83c/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20220722155255-886fb9371eb4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.1.0/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sync v0.7.0 h1:YsImfSBoP9QPYL0xyKJPq0gcaJdG3rInoqxTWbfQu9M=
-golang.org/x/sync v0.7.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
 golang.org/x/sync v0.8.0 h1:3NFvSEYkUoMifnESzZl15y791HH1qU2xm6eCJU5ZPXQ=
 golang.org/x/sync v0.8.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
 golang.org/x/sys v0.0.0-20180810173357-98c5dad5d1a0/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
@@ -990,20 +945,15 @@ golang.org/x/sys v0.0.0-20181029174526-d69651ed3497/go.mod h1:STP8DvDyc/dI5b8T5h
 golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20190316082340-a2f829d7f35f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20190412213103-97732733099d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20190904154756-749cb33beabd/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20190916202348-b4ddaad3f8a3/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20191005200804-aed5e4c7ecf9/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20191026070338-33540a1f6037/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20191120155948-bd437916bb0e/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20200124204421-9fbb57f87de9/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20200217220822-9197077df867/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20200323222414-85ca7c5b95cd/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20200602225109-6fdc65e7d980/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20200728102440-3e129f6d46b1/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20200930185726-fdedc70b468f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20201204225414-ed752295db88/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210112080510-489259a85091/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20210303074136-134d130e1a04/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20210330210617-4fbd30eecc44/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20210423082822-04245dca01da/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
@@ -1017,6 +967,7 @@ golang.org/x/sys v0.0.0-20220715151400-c0bba94af5f8/go.mod h1:oPkhp1MJrh7nUepCBc
 golang.org/x/sys v0.0.0-20220722155257-8c9f86f7a55f/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.0.0-20220811171246-fbc7d0a398ab/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.0.0-20220906165534-d0df966e6959/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.1.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.2.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.3.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.5.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
@@ -1028,11 +979,7 @@ golang.org/x/sys v0.10.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.11.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.12.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.16.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
-golang.org/x/sys v0.17.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
-golang.org/x/sys v0.18.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
 golang.org/x/sys v0.20.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
-golang.org/x/sys v0.22.0 h1:RI27ohtqKCnwULzJLqkv897zojh5/DwS/ENaMzUOaWI=
-golang.org/x/sys v0.22.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
 golang.org/x/sys v0.24.0 h1:Twjiwq9dn6R1fQcyiK+wQyHWfaz/BJB+YIpzU/Cv3Xg=
 golang.org/x/sys v0.24.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
 golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
@@ -1042,14 +989,9 @@ golang.org/x/term v0.3.0/go.mod h1:q750SLmJuPmVoN1blW3UFBPREJfb1KmY3vwxfr+nFDA=
 golang.org/x/term v0.5.0/go.mod h1:jMB1sMXY+tzblOD4FWmEbocvup2/aLOaQEp7JmGp78k=
 golang.org/x/term v0.7.0/go.mod h1:P32HKFT3hSsZrRxla30E9HqToFYAQPCMs/zFMBUFqPY=
 golang.org/x/term v0.8.0/go.mod h1:xPskH00ivmX89bAKVGSKKtLOWNx2+17Eiy94tnKShWo=
-golang.org/x/term v0.10.0/go.mod h1:lpqdcUyK/oCiQxvxVrppt5ggO2KCZ5QblwqPnfZ6d5o=
 golang.org/x/term v0.11.0/go.mod h1:zC9APTIj3jG3FdV/Ons+XE1riIZXG4aZ4GTHiPZJPIU=
 golang.org/x/term v0.16.0/go.mod h1:yn7UURbUtPyrVJPGPq404EukNFxcm/foM+bV/bfcDsY=
-golang.org/x/term v0.17.0/go.mod h1:lLRBjIVuehSbZlaOtGMbcMncT+aqLLLmKrsjNrUguwk=
-golang.org/x/term v0.18.0/go.mod h1:ILwASektA3OnRv7amZ1xhE/KTR+u50pbXfZ03+6Nx58=
 golang.org/x/term v0.20.0/go.mod h1:8UkIAJTvZgivsXaD6/pH6U9ecQzZ45awqEOzuCvwpFY=
-golang.org/x/term v0.21.0 h1:WVXCp+/EBEHOj53Rvu+7KiT/iElMrO8ACK16SMZ3jaA=
-golang.org/x/term v0.21.0/go.mod h1:ooXLefLobQVslOqselCNF4SxFAaoS6KujMbsGzSDmX0=
 golang.org/x/term v0.23.0 h1:F6D4vR+EHoL9/sWAWgAR1H2DcHr4PareCbAaCo1RpuU=
 golang.org/x/term v0.23.0/go.mod h1:DgV24QBUrK6jhZXl+20l6UWznPlwAHm1Q1mGHtydmSk=
 golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
@@ -1061,11 +1003,8 @@ golang.org/x/text v0.4.0/go.mod h1:mrYo+phRRbMaCq/xk9113O4dZlRixOauAjOtrjsXDZ8=
 golang.org/x/text v0.5.0/go.mod h1:mrYo+phRRbMaCq/xk9113O4dZlRixOauAjOtrjsXDZ8=
 golang.org/x/text v0.7.0/go.mod h1:mrYo+phRRbMaCq/xk9113O4dZlRixOauAjOtrjsXDZ8=
 golang.org/x/text v0.9.0/go.mod h1:e1OnstbJyHTd6l/uOt8jFFHp6TRDWZR/bV3emEE/zU8=
-golang.org/x/text v0.11.0/go.mod h1:TvPlkZtksWOMsz7fbANvkp4WM8x/WCo/om8BMLbz+aE=
 golang.org/x/text v0.12.0/go.mod h1:TvPlkZtksWOMsz7fbANvkp4WM8x/WCo/om8BMLbz+aE=
 golang.org/x/text v0.14.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
-golang.org/x/text v0.16.0 h1:a94ExnEXNtEwYLGJSIUxnWoxoRz/ZcCsV63ROupILh4=
-golang.org/x/text v0.16.0/go.mod h1:GhwF1Be+LQoKShO3cGOHzqOgRrGaYc9AvblQOmPVHnI=
 golang.org/x/text v0.17.0 h1:XtiM5bkSOt+ewxlOE/aE/AKEHibwj/6gvWMl9Rsh0Qc=
 golang.org/x/text v0.17.0/go.mod h1:BuEKDfySbSR4drPmRPG/7iBdf8hvFMuRexcpahXilzY=
 golang.org/x/time v0.0.0-20180412165947-fbb02b2291d2/go.mod h1:tRJNPiyCQ0inRvYxbN9jk5I+vvW/OXSQhTDSoE431IQ=
@@ -1087,14 +1026,11 @@ golang.org/x/tools v0.0.0-20191029190741-b9c20aec41a5/go.mod h1:b+2E5dAYhXwXZwtn
 golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
 golang.org/x/tools v0.0.0-20200130002326-2f3ba24bd6e7/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
 golang.org/x/tools v0.0.0-20200619180055-7c47624df98f/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
-golang.org/x/tools v0.0.0-20201224043029-2b0845dc783e/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
 golang.org/x/tools v0.0.0-20210106214847-113979e3529a/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
 golang.org/x/tools v0.1.5/go.mod h1:o0xws9oXOQQZyjljx8fwUC0k7L1pTE6eaCbjGeHmOkk=
 golang.org/x/tools v0.1.12/go.mod h1:hNGJHUnrk76NpqgfD5Aqm5Crs+Hm0VOH/i9J2+nxYbc=
 golang.org/x/tools v0.4.0/go.mod h1:UE5sM2OK9E/d67R0ANs2xJizIymRP5gJU295PvKXxjQ=
 golang.org/x/tools v0.6.0/go.mod h1:Xwgl3UAJ/d3gWutnCtw505GrjyAbvKui8lOU390QaIU=
-golang.org/x/tools v0.22.0 h1:gqSGLZqv+AI9lIQzniJ0nZDRG5GBPsSi+DRNHWNz6yA=
-golang.org/x/tools v0.22.0/go.mod h1:aCwcsjqvq7Yqt6TNyX7QMU2enbQ/Gt0bo6krSeEri+c=
 golang.org/x/tools v0.24.0 h1:J1shsA93PJUEVaUSaay7UXAyE8aimq3GW0pjlolpa24=
 golang.org/x/tools v0.24.0/go.mod h1:YhNqVBIfWHdzvTLs0d8LCuMhkKUgSUKldakyV7W/WDQ=
 golang.org/x/xerrors v0.0.0-20190717185122-a985d3407aa7/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
@@ -1107,11 +1043,13 @@ golang.zx2c4.com/wireguard v0.0.0-20220703234212-c31a7b1ab478 h1:vDy//hdR+GnROE3
 golang.zx2c4.com/wireguard v0.0.0-20220703234212-c31a7b1ab478/go.mod h1:bVQfyl2sCM/QIIGHpWbFGfHPuDvqnCNkT6MQLTCjO/U=
 golang.zx2c4.com/wireguard/windows v0.5.3 h1:On6j2Rpn3OEMXqBq00QEDC7bWSZrPIHKIus8eIuExIE=
 golang.zx2c4.com/wireguard/windows v0.5.3/go.mod h1:9TEe8TJmtwyQebdFwAkEWOPr3prrtqm+REGFifP60hI=
-gonum.org/v1/gonum v0.13.0 h1:a0T3bh+7fhRyqeNbiC3qVHYmkiQgit3wnNan/2c0HMM=
-gonum.org/v1/gonum v0.13.0/go.mod h1:/WPYRckkfWrhWefxyYTfrTtQR0KH4iyHNuzxqXAKyAU=
+gonum.org/v1/gonum v0.15.0 h1:2lYxjRbTYyxkJxlhC+LvJIx3SsANPdRybu1tGj9/OrQ=
+gonum.org/v1/gonum v0.15.0/go.mod h1:xzZVBJBtS+Mz4q0Yl2LJTk+OxOg4jiXZ7qBoM0uISGo=
 google.golang.org/api v0.0.0-20180910000450-7ca32eb868bf/go.mod h1:4mhQ8q/RsB7i+udVvVy5NUi08OU8ZlA0gRVgrF7VFY0=
 google.golang.org/api v0.0.0-20181030000543-1d582fd0359e/go.mod h1:4mhQ8q/RsB7i+udVvVy5NUi08OU8ZlA0gRVgrF7VFY0=
 google.golang.org/api v0.1.0/go.mod h1:UGEZY7KEX120AnNLIHFMKIo4obdJhkp2tPbaPlQx13Y=
+google.golang.org/api v0.180.0 h1:M2D87Yo0rGBPWpo1orwfCLehUUL6E7/TYe5gvMQWDh4=
+google.golang.org/api v0.180.0/go.mod h1:51AiyoEg1MJPSZ9zvklA8VnRILPXxn1iVen9v25XHAE=
 google.golang.org/appengine v1.1.0/go.mod h1:EbEs0AVv82hx2wNQdGPgUI5lhzA/G0D9YwlJXL52JkM=
 google.golang.org/appengine v1.2.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
 google.golang.org/appengine v1.3.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
@@ -1123,8 +1061,12 @@ google.golang.org/genproto v0.0.0-20181202183823-bd91e49a0898/go.mod h1:7Ep/1NZk
 google.golang.org/genproto v0.0.0-20190306203927-b5d61aea6440/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
 google.golang.org/genproto v0.0.0-20190819201941-24fa4b261c55/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
 google.golang.org/genproto v0.0.0-20200526211855-cb27e3aa2013/go.mod h1:NbSheEEYHJ7i3ixzK3sjbqSGDJWnxyFXZblF3eUsNvo=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20240528184218-531527333157 h1:Zy9XzmMEflZ/MAaA7vNcoebnRAld7FsPW1EeBB7V0m8=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20240528184218-531527333157/go.mod h1:EfXuqaE1J41VCDicxHzUDm+8rk+7ZdXzHV0IhO/I6s0=
+google.golang.org/genproto v0.0.0-20240401170217-c3f982113cda h1:wu/KJm9KJwpfHWhkkZGohVC6KRrc1oJNr4jwtQMOQXw=
+google.golang.org/genproto v0.0.0-20240401170217-c3f982113cda/go.mod h1:g2LLCvCeCSir/JJSWosk19BR4NVxGqHUC6rxIRsd7Aw=
+google.golang.org/genproto/googleapis/api v0.0.0-20240617180043-68d350f18fd4 h1:MuYw1wJzT+ZkybKfaOXKp5hJiZDn2iHaXRw0mRYdHSc=
+google.golang.org/genproto/googleapis/api v0.0.0-20240617180043-68d350f18fd4/go.mod h1:px9SlOOZBg1wM1zdnr8jEL4CNGUBZ+ZKYtNPApNQc4c=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20240617180043-68d350f18fd4 h1:Di6ANFilr+S60a4S61ZM00vLdw0IrQOSMS2/6mrnOU0=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20240617180043-68d350f18fd4/go.mod h1:Ue6ibwXGpU+dqIcODieyLOcgj7z8+IcskoNIgZxtrFY=
 google.golang.org/grpc v1.14.0/go.mod h1:yo6s7OP7yaDglbqo1J04qKzAhqBH6lvTonzMVmEdcZw=
 google.golang.org/grpc v1.16.0/go.mod h1:0JHn/cJsOMiMfNA9+DeHDlAU7KAAB5GDlYFpa9MZMio=
 google.golang.org/grpc v1.17.0/go.mod h1:6QZJwpn2B+Zp71q/5VxRsJ6NXXVCE5NRUHRo+f3cWCs=
@@ -1144,8 +1086,6 @@ google.golang.org/protobuf v1.22.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2
 google.golang.org/protobuf v1.23.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
 google.golang.org/protobuf v1.23.1-0.20200526195155-81db48ad09cc/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
 google.golang.org/protobuf v1.25.0/go.mod h1:9JNX74DMeImyA3h4bdi1ymwjUzf21/xIlbajtzgsN7c=
-google.golang.org/protobuf v1.26.0-rc.1/go.mod h1:jlhhOSvTdKEhbULTjvd4ARK9grFBp09yW+WbY/TyQbw=
-google.golang.org/protobuf v1.26.0/go.mod h1:9q0QmTI4eRPtz6boOQmLYwt+qCgq0jsYwAQnmE0givc=
 google.golang.org/protobuf v1.34.2 h1:6xV6lTsCfpGD21XK49h7MhtcApnLqkfYgPcdHftf6hg=
 google.golang.org/protobuf v1.34.2/go.mod h1:qYOHts0dSfpeUzUFpOMr/WGzszTmLH+DiWniOlNbLDw=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
@@ -1157,13 +1097,11 @@ gopkg.in/errgo.v2 v2.1.0/go.mod h1:hNsd1EY+bozCKY1Ytp96fpM3vjJbqLJn88ws8XvfDNI=
 gopkg.in/fsnotify.v1 v1.4.7 h1:xOHLXZwVvI9hhs+cLKq5+I5onOuwQLhQwiu63xxlHs4=
 gopkg.in/fsnotify.v1 v1.4.7/go.mod h1:Tz8NjZHkW78fSQdbUxIjBTcgA1z1m8ZHf0WmKUhAMys=
 gopkg.in/inf.v0 v0.9.1/go.mod h1:cWUDdTG/fYaXco+Dcufb5Vnc6Gp2YChqWtbxRZE0mXw=
-gopkg.in/op/go-logging.v1 v1.0.0-20160211212156-b2cb9fa56473/go.mod h1:N1eN2tsCx0Ydtgjl4cqmbRCsY4/+z4cYDeqwZTk6zog=
 gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7 h1:uRGJdciOHaEIrze2W8Q3AKkepLTh2hOroT7a+7czfdQ=
 gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7/go.mod h1:dt/ZhP58zS4L8KSrWDmTeBkI65Dw0HsyUHuEVlX15mw=
 gopkg.in/yaml.v1 v1.0.0-20140924161607-9f9df34309c0/go.mod h1:WDnlLJ4WF5VGsH/HVa3CI79GS0ol3YnhVnKP89i0kNg=
 gopkg.in/yaml.v2 v2.2.1/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
-gopkg.in/yaml.v2 v2.2.4/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v2 v2.2.8/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v2 v2.3.0/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v2 v2.4.0 h1:D8xgwECY7CYvx+Y2n4sBz93Jn9JRvxdiyyo8CTfuKaY=
@@ -1182,8 +1120,6 @@ honnef.co/go/tools v0.0.0-20190523083050-ea95bdfd59fc/go.mod h1:rf3lG4BRIbNafJWh
 honnef.co/go/tools v0.0.1-2019.2.3/go.mod h1:a3bituU0lyd329TUQxRnasdCoJDkEUEAqEt0JzvZhAg=
 howett.net/plist v1.0.0 h1:7CrbWYbPPO/PyNy38b2EB/+gYbjCe2DXBxgtOOZbSQM=
 howett.net/plist v1.0.0/go.mod h1:lqaXoTrLY4hg8tnEzNru53gicrbv7rrk+2xJA/7hw9g=
-lukechampine.com/blake3 v1.2.1 h1:YuqqRuaqsGV71BV/nm9xlI0MKUv4QC54jQnBChWbGnI=
-lukechampine.com/blake3 v1.2.1/go.mod h1:0OFRp7fBtAylGVCO40o87sbupkyIGgbpv1+M1k1LM6k=
 lukechampine.com/blake3 v1.3.0 h1:sJ3XhFINmHSrYCgl958hscfIa3bw8x4DqMP3u1YvoYE=
 lukechampine.com/blake3 v1.3.0/go.mod h1:0OFRp7fBtAylGVCO40o87sbupkyIGgbpv1+M1k1LM6k=
 oras.land/oras-go/v2 v2.5.0 h1:o8Me9kLY74Vp5uw07QXPiitjsw7qNXi8Twd+19Zf02c=
diff --git a/pkg/grpc/backend.go b/pkg/grpc/backend.go
index 596a7589..85c9e5bc 100644
--- a/pkg/grpc/backend.go
+++ b/pkg/grpc/backend.go
@@ -3,7 +3,6 @@ package grpc
 import (
 	"context"
 
-	"github.com/mudler/LocalAI/core/schema"
 	pb "github.com/mudler/LocalAI/pkg/grpc/proto"
 	"google.golang.org/grpc"
 )
@@ -18,10 +17,10 @@ func NewClient(address string, parallel bool, wd WatchDog, enableWatchDog bool)
 	if bc, ok := embeds[address]; ok {
 		return bc
 	}
-	return NewGrpcClient(address, parallel, wd, enableWatchDog)
+	return buildClient(address, parallel, wd, enableWatchDog)
 }
 
-func NewGrpcClient(address string, parallel bool, wd WatchDog, enableWatchDog bool) Backend {
+func buildClient(address string, parallel bool, wd WatchDog, enableWatchDog bool) Backend {
 	if !enableWatchDog {
 		wd = nil
 	}
@@ -41,7 +40,8 @@ type Backend interface {
 	PredictStream(ctx context.Context, in *pb.PredictOptions, f func(s []byte), opts ...grpc.CallOption) error
 	GenerateImage(ctx context.Context, in *pb.GenerateImageRequest, opts ...grpc.CallOption) (*pb.Result, error)
 	TTS(ctx context.Context, in *pb.TTSRequest, opts ...grpc.CallOption) (*pb.Result, error)
-	AudioTranscription(ctx context.Context, in *pb.TranscriptRequest, opts ...grpc.CallOption) (*schema.TranscriptionResult, error)
+	SoundGeneration(ctx context.Context, in *pb.SoundGenerationRequest, opts ...grpc.CallOption) (*pb.Result, error)
+	AudioTranscription(ctx context.Context, in *pb.TranscriptRequest, opts ...grpc.CallOption) (*pb.TranscriptResult, error)
 	TokenizeString(ctx context.Context, in *pb.PredictOptions, opts ...grpc.CallOption) (*pb.TokenizationResponse, error)
 	Status(ctx context.Context) (*pb.StatusResponse, error)
 
diff --git a/pkg/grpc/base/base.go b/pkg/grpc/base/base.go
index 515022ec..95dca561 100644
--- a/pkg/grpc/base/base.go
+++ b/pkg/grpc/base/base.go
@@ -6,7 +6,6 @@ import (
 	"fmt"
 	"os"
 
-	"github.com/mudler/LocalAI/core/schema"
 	pb "github.com/mudler/LocalAI/pkg/grpc/proto"
 	gopsutil "github.com/shirou/gopsutil/v3/process"
 )
@@ -53,14 +52,18 @@ func (llm *Base) GenerateImage(*pb.GenerateImageRequest) error {
 	return fmt.Errorf("unimplemented")
 }
 
-func (llm *Base) AudioTranscription(*pb.TranscriptRequest) (schema.TranscriptionResult, error) {
-	return schema.TranscriptionResult{}, fmt.Errorf("unimplemented")
+func (llm *Base) AudioTranscription(*pb.TranscriptRequest) (pb.TranscriptResult, error) {
+	return pb.TranscriptResult{}, fmt.Errorf("unimplemented")
 }
 
 func (llm *Base) TTS(*pb.TTSRequest) error {
 	return fmt.Errorf("unimplemented")
 }
 
+func (llm *Base) SoundGeneration(*pb.SoundGenerationRequest) error {
+	return fmt.Errorf("unimplemented")
+}
+
 func (llm *Base) TokenizeString(opts *pb.PredictOptions) (pb.TokenizationResponse, error) {
 	return pb.TokenizationResponse{}, fmt.Errorf("unimplemented")
 }
diff --git a/pkg/grpc/client.go b/pkg/grpc/client.go
index cfae5875..032c9c00 100644
--- a/pkg/grpc/client.go
+++ b/pkg/grpc/client.go
@@ -7,7 +7,6 @@ import (
 	"sync"
 	"time"
 
-	"github.com/mudler/LocalAI/core/schema"
 	pb "github.com/mudler/LocalAI/pkg/grpc/proto"
 	"google.golang.org/grpc"
 	"google.golang.org/grpc/credentials/insecure"
@@ -39,6 +38,18 @@ func (c *Client) setBusy(v bool) {
 	c.Unlock()
 }
 
+func (c *Client) wdMark() {
+	if c.wd != nil {
+		c.wd.Mark(c.address)
+	}
+}
+
+func (c *Client) wdUnMark() {
+	if c.wd != nil {
+		c.wd.UnMark(c.address)
+	}
+}
+
 func (c *Client) HealthCheck(ctx context.Context) (bool, error) {
 	if !c.parallel {
 		c.opMutex.Lock()
@@ -76,10 +87,8 @@ func (c *Client) Embeddings(ctx context.Context, in *pb.PredictOptions, opts ...
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
-	if c.wd != nil {
-		c.wd.Mark(c.address)
-		defer c.wd.UnMark(c.address)
-	}
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
@@ -97,10 +106,8 @@ func (c *Client) Predict(ctx context.Context, in *pb.PredictOptions, opts ...grp
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
-	if c.wd != nil {
-		c.wd.Mark(c.address)
-		defer c.wd.UnMark(c.address)
-	}
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
@@ -118,10 +125,8 @@ func (c *Client) LoadModel(ctx context.Context, in *pb.ModelOptions, opts ...grp
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
-	if c.wd != nil {
-		c.wd.Mark(c.address)
-		defer c.wd.UnMark(c.address)
-	}
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
@@ -138,10 +143,8 @@ func (c *Client) PredictStream(ctx context.Context, in *pb.PredictOptions, f fun
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
-	if c.wd != nil {
-		c.wd.Mark(c.address)
-		defer c.wd.UnMark(c.address)
-	}
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return err
@@ -177,10 +180,8 @@ func (c *Client) GenerateImage(ctx context.Context, in *pb.GenerateImageRequest,
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
-	if c.wd != nil {
-		c.wd.Mark(c.address)
-		defer c.wd.UnMark(c.address)
-	}
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
@@ -197,10 +198,8 @@ func (c *Client) TTS(ctx context.Context, in *pb.TTSRequest, opts ...grpc.CallOp
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
-	if c.wd != nil {
-		c.wd.Mark(c.address)
-		defer c.wd.UnMark(c.address)
-	}
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
@@ -210,44 +209,40 @@ func (c *Client) TTS(ctx context.Context, in *pb.TTSRequest, opts ...grpc.CallOp
 	return client.TTS(ctx, in, opts...)
 }
 
-func (c *Client) AudioTranscription(ctx context.Context, in *pb.TranscriptRequest, opts ...grpc.CallOption) (*schema.TranscriptionResult, error) {
+func (c *Client) SoundGeneration(ctx context.Context, in *pb.SoundGenerationRequest, opts ...grpc.CallOption) (*pb.Result, error) {
 	if !c.parallel {
 		c.opMutex.Lock()
 		defer c.opMutex.Unlock()
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
-	if c.wd != nil {
-		c.wd.Mark(c.address)
-		defer c.wd.UnMark(c.address)
-	}
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
 	}
 	defer conn.Close()
 	client := pb.NewBackendClient(conn)
-	res, err := client.AudioTranscription(ctx, in, opts...)
+	return client.SoundGeneration(ctx, in, opts...)
+}
+
+func (c *Client) AudioTranscription(ctx context.Context, in *pb.TranscriptRequest, opts ...grpc.CallOption) (*pb.TranscriptResult, error) {
+	if !c.parallel {
+		c.opMutex.Lock()
+		defer c.opMutex.Unlock()
+	}
+	c.setBusy(true)
+	defer c.setBusy(false)
+	c.wdMark()
+	defer c.wdUnMark()
+	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
 	}
-	tresult := &schema.TranscriptionResult{}
-	for _, s := range res.Segments {
-		tks := []int{}
-		for _, t := range s.Tokens {
-			tks = append(tks, int(t))
-		}
-		tresult.Segments = append(tresult.Segments,
-			schema.Segment{
-				Text:   s.Text,
-				Id:     int(s.Id),
-				Start:  time.Duration(s.Start),
-				End:    time.Duration(s.End),
-				Tokens: tks,
-			})
-	}
-	tresult.Text = res.Text
-	return tresult, err
+	defer conn.Close()
+	client := pb.NewBackendClient(conn)
+	return client.AudioTranscription(ctx, in, opts...)
 }
 
 func (c *Client) TokenizeString(ctx context.Context, in *pb.PredictOptions, opts ...grpc.CallOption) (*pb.TokenizationResponse, error) {
@@ -257,10 +252,8 @@ func (c *Client) TokenizeString(ctx context.Context, in *pb.PredictOptions, opts
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
-	if c.wd != nil {
-		c.wd.Mark(c.address)
-		defer c.wd.UnMark(c.address)
-	}
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
@@ -299,6 +292,8 @@ func (c *Client) StoresSet(ctx context.Context, in *pb.StoresSetOptions, opts ..
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
@@ -313,6 +308,8 @@ func (c *Client) StoresDelete(ctx context.Context, in *pb.StoresDeleteOptions, o
 		c.opMutex.Lock()
 		defer c.opMutex.Unlock()
 	}
+	c.wdMark()
+	defer c.wdUnMark()
 	c.setBusy(true)
 	defer c.setBusy(false)
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
@@ -331,6 +328,8 @@ func (c *Client) StoresGet(ctx context.Context, in *pb.StoresGetOptions, opts ..
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
@@ -347,6 +346,8 @@ func (c *Client) StoresFind(ctx context.Context, in *pb.StoresFindOptions, opts
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
@@ -363,6 +364,8 @@ func (c *Client) Rerank(ctx context.Context, in *pb.RerankRequest, opts ...grpc.
 	}
 	c.setBusy(true)
 	defer c.setBusy(false)
+	c.wdMark()
+	defer c.wdUnMark()
 	conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
 	if err != nil {
 		return nil, err
diff --git a/pkg/grpc/embed.go b/pkg/grpc/embed.go
index 2b776b39..3155ff59 100644
--- a/pkg/grpc/embed.go
+++ b/pkg/grpc/embed.go
@@ -2,9 +2,7 @@ package grpc
 
 import (
 	"context"
-	"time"
 
-	"github.com/mudler/LocalAI/core/schema"
 	pb "github.com/mudler/LocalAI/pkg/grpc/proto"
 	"google.golang.org/grpc"
 	"google.golang.org/grpc/metadata"
@@ -53,28 +51,12 @@ func (e *embedBackend) TTS(ctx context.Context, in *pb.TTSRequest, opts ...grpc.
 	return e.s.TTS(ctx, in)
 }
 
-func (e *embedBackend) AudioTranscription(ctx context.Context, in *pb.TranscriptRequest, opts ...grpc.CallOption) (*schema.TranscriptionResult, error) {
-	r, err := e.s.AudioTranscription(ctx, in)
-	if err != nil {
-		return nil, err
-	}
-	tr := &schema.TranscriptionResult{}
-	for _, s := range r.Segments {
-		var tks []int
-		for _, t := range s.Tokens {
-			tks = append(tks, int(t))
-		}
-		tr.Segments = append(tr.Segments,
-			schema.Segment{
-				Text:   s.Text,
-				Id:     int(s.Id),
-				Start:  time.Duration(s.Start),
-				End:    time.Duration(s.End),
-				Tokens: tks,
-			})
-	}
-	tr.Text = r.Text
-	return tr, err
+func (e *embedBackend) SoundGeneration(ctx context.Context, in *pb.SoundGenerationRequest, opts ...grpc.CallOption) (*pb.Result, error) {
+	return e.s.SoundGeneration(ctx, in)
+}
+
+func (e *embedBackend) AudioTranscription(ctx context.Context, in *pb.TranscriptRequest, opts ...grpc.CallOption) (*pb.TranscriptResult, error) {
+	return e.s.AudioTranscription(ctx, in)
 }
 
 func (e *embedBackend) TokenizeString(ctx context.Context, in *pb.PredictOptions, opts ...grpc.CallOption) (*pb.TokenizationResponse, error) {
diff --git a/pkg/grpc/interface.go b/pkg/grpc/interface.go
index 313c8ff5..97b958cc 100644
--- a/pkg/grpc/interface.go
+++ b/pkg/grpc/interface.go
@@ -1,7 +1,6 @@
 package grpc
 
 import (
-	"github.com/mudler/LocalAI/core/schema"
 	pb "github.com/mudler/LocalAI/pkg/grpc/proto"
 )
 
@@ -15,8 +14,9 @@ type LLM interface {
 	Load(*pb.ModelOptions) error
 	Embeddings(*pb.PredictOptions) ([]float32, error)
 	GenerateImage(*pb.GenerateImageRequest) error
-	AudioTranscription(*pb.TranscriptRequest) (schema.TranscriptionResult, error)
+	AudioTranscription(*pb.TranscriptRequest) (pb.TranscriptResult, error)
 	TTS(*pb.TTSRequest) error
+	SoundGeneration(*pb.SoundGenerationRequest) error
 	TokenizeString(*pb.PredictOptions) (pb.TokenizationResponse, error)
 	Status() (pb.StatusResponse, error)
 
diff --git a/pkg/grpc/server.go b/pkg/grpc/server.go
index 784aac7f..0e602a42 100644
--- a/pkg/grpc/server.go
+++ b/pkg/grpc/server.go
@@ -84,7 +84,19 @@ func (s *server) TTS(ctx context.Context, in *pb.TTSRequest) (*pb.Result, error)
 	if err != nil {
 		return &pb.Result{Message: fmt.Sprintf("Error generating audio: %s", err.Error()), Success: false}, err
 	}
-	return &pb.Result{Message: "Audio generated", Success: true}, nil
+	return &pb.Result{Message: "TTS audio generated", Success: true}, nil
+}
+
+func (s *server) SoundGeneration(ctx context.Context, in *pb.SoundGenerationRequest) (*pb.Result, error) {
+	if s.llm.Locking() {
+		s.llm.Lock()
+		defer s.llm.Unlock()
+	}
+	err := s.llm.SoundGeneration(in)
+	if err != nil {
+		return &pb.Result{Message: fmt.Sprintf("Error generating audio: %s", err.Error()), Success: false}, err
+	}
+	return &pb.Result{Message: "Sound Generation audio generated", Success: true}, nil
 }
 
 func (s *server) AudioTranscription(ctx context.Context, in *pb.TranscriptRequest) (*pb.TranscriptResult, error) {
diff --git a/pkg/model/initializers.go b/pkg/model/initializers.go
index 11980f03..3d2255cc 100644
--- a/pkg/model/initializers.go
+++ b/pkg/model/initializers.go
@@ -80,6 +80,9 @@ ENTRY:
 		if e.IsDir() {
 			continue
 		}
+		if strings.HasSuffix(e.Name(), ".log") {
+			continue
+		}
 
 		// Skip the llama.cpp variants if we are autoDetecting
 		// But we always load the fallback variant if it exists
@@ -265,12 +268,12 @@ func selectGRPCProcess(backend, assetDir string, f16 bool) string {
 
 // starts the grpcModelProcess for the backend, and returns a grpc client
 // It also loads the model
-func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string) (ModelAddress, error) {
-	return func(modelName, modelFile string) (ModelAddress, error) {
+func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string) (*Model, error) {
+	return func(modelName, modelFile string) (*Model, error) {
 
 		log.Debug().Msgf("Loading Model %s with gRPC (file: %s) (backend: %s): %+v", modelName, modelFile, backend, *o)
 
-		var client ModelAddress
+		var client *Model
 
 		getFreeAddress := func() (string, error) {
 			port, err := freeport.GetFreePort()
@@ -294,27 +297,30 @@ func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string
 		if uri, ok := o.externalBackends[backend]; ok {
 			log.Debug().Msgf("Loading external backend: %s", uri)
 			// check if uri is a file or a address
-			if _, err := os.Stat(uri); err == nil {
+			if fi, err := os.Stat(uri); err == nil {
+				log.Debug().Msgf("external backend is file: %+v", fi)
 				serverAddress, err := getFreeAddress()
 				if err != nil {
-					return "", fmt.Errorf("failed allocating free ports: %s", err.Error())
+					return nil, fmt.Errorf("failed allocating free ports: %s", err.Error())
 				}
 				// Make sure the process is executable
 				if err := ml.startProcess(uri, o.model, serverAddress); err != nil {
-					return "", err
+					log.Error().Err(err).Str("path", uri).Msg("failed to launch ")
+					return nil, err
 				}
 
 				log.Debug().Msgf("GRPC Service Started")
 
-				client = ModelAddress(serverAddress)
+				client = NewModel(serverAddress)
 			} else {
+				log.Debug().Msg("external backend is uri")
 				// address
-				client = ModelAddress(uri)
+				client = NewModel(uri)
 			}
 		} else {
 			grpcProcess := backendPath(o.assetDir, backend)
 			if err := utils.VerifyPath(grpcProcess, o.assetDir); err != nil {
-				return "", fmt.Errorf("grpc process not found in assetdir: %s", err.Error())
+				return nil, fmt.Errorf("grpc process not found in assetdir: %s", err.Error())
 			}
 
 			if autoDetect {
@@ -326,12 +332,12 @@ func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string
 
 			// Check if the file exists
 			if _, err := os.Stat(grpcProcess); os.IsNotExist(err) {
-				return "", fmt.Errorf("grpc process not found: %s. some backends(stablediffusion, tts) require LocalAI compiled with GO_TAGS", grpcProcess)
+				return nil, fmt.Errorf("grpc process not found: %s. some backends(stablediffusion, tts) require LocalAI compiled with GO_TAGS", grpcProcess)
 			}
 
 			serverAddress, err := getFreeAddress()
 			if err != nil {
-				return "", fmt.Errorf("failed allocating free ports: %s", err.Error())
+				return nil, fmt.Errorf("failed allocating free ports: %s", err.Error())
 			}
 
 			args := []string{}
@@ -341,12 +347,12 @@ func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string
 
 			// Make sure the process is executable in any circumstance
 			if err := ml.startProcess(grpcProcess, o.model, serverAddress, args...); err != nil {
-				return "", err
+				return nil, err
 			}
 
 			log.Debug().Msgf("GRPC Service Started")
 
-			client = ModelAddress(serverAddress)
+			client = NewModel(serverAddress)
 		}
 
 		// Wait for the service to start up
@@ -366,7 +372,7 @@ func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string
 
 		if !ready {
 			log.Debug().Msgf("GRPC Service NOT ready")
-			return "", fmt.Errorf("grpc service not ready")
+			return nil, fmt.Errorf("grpc service not ready")
 		}
 
 		options := *o.gRPCOptions
@@ -377,25 +383,18 @@ func (ml *ModelLoader) grpcModel(backend string, o *Options) func(string, string
 
 		res, err := client.GRPC(o.parallelRequests, ml.wd).LoadModel(o.context, &options)
 		if err != nil {
-			return "", fmt.Errorf("could not load model: %w", err)
+			return nil, fmt.Errorf("could not load model: %w", err)
 		}
 		if !res.Success {
-			return "", fmt.Errorf("could not load model (no success): %s", res.Message)
+			return nil, fmt.Errorf("could not load model (no success): %s", res.Message)
 		}
 
 		return client, nil
 	}
 }
 
-func (ml *ModelLoader) resolveAddress(addr ModelAddress, parallel bool) (grpc.Backend, error) {
-	if parallel {
-		return addr.GRPC(parallel, ml.wd), nil
-	}
-
-	if _, ok := ml.grpcClients[string(addr)]; !ok {
-		ml.grpcClients[string(addr)] = addr.GRPC(parallel, ml.wd)
-	}
-	return ml.grpcClients[string(addr)], nil
+func (ml *ModelLoader) ListAvailableBackends(assetdir string) ([]string, error) {
+	return backendsInAssetDir(assetdir)
 }
 
 func (ml *ModelLoader) BackendLoader(opts ...Option) (client grpc.Backend, err error) {
@@ -422,7 +421,6 @@ func (ml *ModelLoader) BackendLoader(opts ...Option) (client grpc.Backend, err e
 			log.Error().Err(err).Str("keptModel", o.model).Msg("error while shutting down all backends except for the keptModel")
 			return nil, err
 		}
-
 	}
 
 	var backendToConsume string
@@ -435,26 +433,28 @@ func (ml *ModelLoader) BackendLoader(opts ...Option) (client grpc.Backend, err e
 		backendToConsume = backend
 	}
 
-	addr, err := ml.LoadModel(o.model, ml.grpcModel(backendToConsume, o))
+	model, err := ml.LoadModel(o.model, ml.grpcModel(backendToConsume, o))
 	if err != nil {
 		return nil, err
 	}
 
-	return ml.resolveAddress(addr, o.parallelRequests)
+	return model.GRPC(o.parallelRequests, ml.wd), nil
 }
 
 func (ml *ModelLoader) GreedyLoader(opts ...Option) (grpc.Backend, error) {
 	o := NewOptions(opts...)
 
 	ml.mu.Lock()
+
 	// Return earlier if we have a model already loaded
 	// (avoid looping through all the backends)
-	if m := ml.CheckIsLoaded(o.model); m != "" {
+	if m := ml.CheckIsLoaded(o.model); m != nil {
 		log.Debug().Msgf("Model '%s' already loaded", o.model)
 		ml.mu.Unlock()
 
-		return ml.resolveAddress(m, o.parallelRequests)
+		return m.GRPC(o.parallelRequests, ml.wd), nil
 	}
+
 	// If we can have only one backend active, kill all the others (except external backends)
 	if o.singleActiveBackend {
 		log.Debug().Msgf("Stopping all backends except '%s'", o.model)
diff --git a/pkg/model/loader.go b/pkg/model/loader.go
index b2570c71..b9865f73 100644
--- a/pkg/model/loader.go
+++ b/pkg/model/loader.go
@@ -7,70 +7,32 @@ import (
 	"path/filepath"
 	"strings"
 	"sync"
+	"time"
 
 	"github.com/mudler/LocalAI/pkg/templates"
 
-	"github.com/mudler/LocalAI/pkg/functions"
-	"github.com/mudler/LocalAI/pkg/grpc"
 	"github.com/mudler/LocalAI/pkg/utils"
 
 	process "github.com/mudler/go-processmanager"
 	"github.com/rs/zerolog/log"
 )
 
-// Rather than pass an interface{} to the prompt template:
-// These are the definitions of all possible variables LocalAI will currently populate for use in a prompt template file
-// Please note: Not all of these are populated on every endpoint - your template should either be tested for each endpoint you map it to, or tolerant of zero values.
-type PromptTemplateData struct {
-	SystemPrompt         string
-	SuppressSystemPrompt bool // used by chat specifically to indicate that SystemPrompt above should be _ignored_
-	Input                string
-	Instruction          string
-	Functions            []functions.Function
-	MessageIndex         int
-}
-
-type ChatMessageTemplateData struct {
-	SystemPrompt string
-	Role         string
-	RoleName     string
-	FunctionName string
-	Content      string
-	MessageIndex int
-	Function     bool
-	FunctionCall interface{}
-	LastMessage  bool
-}
-
 // new idea: what if we declare a struct of these here, and use a loop to check?
 
 // TODO: Split ModelLoader and TemplateLoader? Just to keep things more organized. Left together to share a mutex until I look into that. Would split if we seperate directories for .bin/.yaml and .tmpl
 type ModelLoader struct {
-	ModelPath string
-	mu        sync.Mutex
-	// TODO: this needs generics
-	grpcClients   map[string]grpc.Backend
-	models        map[string]ModelAddress
+	ModelPath     string
+	mu            sync.Mutex
+	models        map[string]*Model
 	grpcProcesses map[string]*process.Process
 	templates     *templates.TemplateCache
 	wd            *WatchDog
 }
 
-type ModelAddress string
-
-func (m ModelAddress) GRPC(parallel bool, wd *WatchDog) grpc.Backend {
-	enableWD := false
-	if wd != nil {
-		enableWD = true
-	}
-	return grpc.NewClient(string(m), parallel, wd, enableWD)
-}
-
 func NewModelLoader(modelPath string) *ModelLoader {
 	nml := &ModelLoader{
 		ModelPath:     modelPath,
-		grpcClients:   make(map[string]grpc.Backend),
-		models:        make(map[string]ModelAddress),
+		models:        make(map[string]*Model),
 		templates:     templates.NewTemplateCache(modelPath),
 		grpcProcesses: make(map[string]*process.Process),
 	}
@@ -107,6 +69,8 @@ var knownModelsNameSuffixToSkip []string = []string{
 	".tar.gz",
 }
 
+const retryTimeout = time.Duration(2 * time.Minute)
+
 func (ml *ModelLoader) ListFilesInModelPath() ([]string, error) {
 	files, err := os.ReadDir(ml.ModelPath)
 	if err != nil {
@@ -141,12 +105,24 @@ FILE:
 	return models, nil
 }
 
-func (ml *ModelLoader) LoadModel(modelName string, loader func(string, string) (ModelAddress, error)) (ModelAddress, error) {
+func (ml *ModelLoader) ListModels() []*Model {
+	ml.mu.Lock()
+	defer ml.mu.Unlock()
+
+	models := []*Model{}
+	for _, model := range ml.models {
+		models = append(models, model)
+	}
+
+	return models
+}
+
+func (ml *ModelLoader) LoadModel(modelName string, loader func(string, string) (*Model, error)) (*Model, error) {
 	ml.mu.Lock()
 	defer ml.mu.Unlock()
 
 	// Check if we already have a loaded model
-	if model := ml.CheckIsLoaded(modelName); model != "" {
+	if model := ml.CheckIsLoaded(modelName); model != nil {
 		return model, nil
 	}
 
@@ -156,18 +132,15 @@ func (ml *ModelLoader) LoadModel(modelName string, loader func(string, string) (
 
 	model, err := loader(modelName, modelFile)
 	if err != nil {
-		return "", err
+		return nil, err
 	}
 
-	// TODO: Add a helper method to iterate all prompt templates associated with a config if and only if it's YAML?
-	// Minor perf loss here until this is fixed, but we initialize on first request
-
-	// // If there is a prompt template, load it
-	// if err := ml.loadTemplateIfExists(modelName); err != nil {
-	// 	return nil, err
-	// }
+	if model == nil {
+		return nil, fmt.Errorf("loader didn't return a model")
+	}
 
 	ml.models[modelName] = model
+
 	return model, nil
 }
 
@@ -175,64 +148,57 @@ func (ml *ModelLoader) ShutdownModel(modelName string) error {
 	ml.mu.Lock()
 	defer ml.mu.Unlock()
 
-	return ml.stopModel(modelName)
-}
-
-func (ml *ModelLoader) stopModel(modelName string) error {
-	defer ml.deleteProcess(modelName)
-	if _, ok := ml.models[modelName]; !ok {
+	_, ok := ml.models[modelName]
+	if !ok {
 		return fmt.Errorf("model %s not found", modelName)
 	}
-	return nil
-	//return ml.deleteProcess(modelName)
+
+	retries := 1
+	for ml.models[modelName].GRPC(false, ml.wd).IsBusy() {
+		log.Debug().Msgf("%s busy. Waiting.", modelName)
+		dur := time.Duration(retries*2) * time.Second
+		if dur > retryTimeout {
+			dur = retryTimeout
+		}
+		time.Sleep(dur)
+		retries++
+	}
+
+	return ml.deleteProcess(modelName)
 }
 
-func (ml *ModelLoader) CheckIsLoaded(s string) ModelAddress {
-	var client grpc.Backend
-	if m, ok := ml.models[s]; ok {
-		log.Debug().Msgf("Model already loaded in memory: %s", s)
-		if c, ok := ml.grpcClients[s]; ok {
-			client = c
-		} else {
-			client = m.GRPC(false, ml.wd)
+func (ml *ModelLoader) CheckIsLoaded(s string) *Model {
+	m, ok := ml.models[s]
+	if !ok {
+		return nil
+	}
+
+	log.Debug().Msgf("Model already loaded in memory: %s", s)
+	client := m.GRPC(false, ml.wd)
+
+	log.Debug().Msgf("Checking model availability (%s)", s)
+	cTimeout, cancel := context.WithTimeout(context.Background(), 2*time.Minute)
+	defer cancel()
+
+	alive, err := client.HealthCheck(cTimeout)
+	if !alive {
+		log.Warn().Msgf("GRPC Model not responding: %s", err.Error())
+		log.Warn().Msgf("Deleting the process in order to recreate it")
+		process, exists := ml.grpcProcesses[s]
+		if !exists {
+			log.Error().Msgf("Process not found for '%s' and the model is not responding anymore !", s)
+			return m
 		}
-		alive, err := client.HealthCheck(context.Background())
-		if !alive {
-			log.Warn().Msgf("GRPC Model not responding: %s", err.Error())
-			log.Warn().Msgf("Deleting the process in order to recreate it")
-			if !ml.grpcProcesses[s].IsAlive() {
-				log.Debug().Msgf("GRPC Process is not responding: %s", s)
-				// stop and delete the process, this forces to re-load the model and re-create again the service
-				err := ml.deleteProcess(s)
-				if err != nil {
-					log.Error().Err(err).Str("process", s).Msg("error stopping process")
-				}
-				return ""
+		if !process.IsAlive() {
+			log.Debug().Msgf("GRPC Process is not responding: %s", s)
+			// stop and delete the process, this forces to re-load the model and re-create again the service
+			err := ml.deleteProcess(s)
+			if err != nil {
+				log.Error().Err(err).Str("process", s).Msg("error stopping process")
 			}
+			return nil
 		}
-
-		return m
 	}
 
-	return ""
-}
-
-const (
-	ChatPromptTemplate templates.TemplateType = iota
-	ChatMessageTemplate
-	CompletionPromptTemplate
-	EditPromptTemplate
-	FunctionsPromptTemplate
-)
-
-func (ml *ModelLoader) EvaluateTemplateForPrompt(templateType templates.TemplateType, templateName string, in PromptTemplateData) (string, error) {
-	// TODO: should this check be improved?
-	if templateType == ChatMessageTemplate {
-		return "", fmt.Errorf("invalid templateType: ChatMessage")
-	}
-	return ml.templates.EvaluateTemplate(templateType, templateName, in)
-}
-
-func (ml *ModelLoader) EvaluateTemplateForChatMessage(templateName string, messageData ChatMessageTemplateData) (string, error) {
-	return ml.templates.EvaluateTemplate(ChatMessageTemplate, templateName, messageData)
+	return m
 }
diff --git a/pkg/model/loader_test.go b/pkg/model/loader_test.go
index 1142ed0c..4621844e 100644
--- a/pkg/model/loader_test.go
+++ b/pkg/model/loader_test.go
@@ -1,197 +1,105 @@
 package model_test
 
 import (
-	. "github.com/mudler/LocalAI/pkg/model"
+	"errors"
+	"os"
+	"path/filepath"
 
+	"github.com/mudler/LocalAI/pkg/model"
 	. "github.com/onsi/ginkgo/v2"
 	. "github.com/onsi/gomega"
 )
 
-const chatML = `<|im_start|>{{if eq .RoleName "assistant"}}assistant{{else if eq .RoleName "system"}}system{{else if eq .RoleName "tool"}}tool{{else if eq .RoleName "user"}}user{{end}}
-{{- if .FunctionCall }}
-<tool_call>
-{{- else if eq .RoleName "tool" }}
-<tool_response>
-{{- end }}
-{{- if .Content}}
-{{.Content }}
-{{- end }}
-{{- if .FunctionCall}}
-{{toJson .FunctionCall}}
-{{- end }}
-{{- if .FunctionCall }}
-</tool_call>
-{{- else if eq .RoleName "tool" }}
-</tool_response>
-{{- end }}<|im_end|>`
+var _ = Describe("ModelLoader", func() {
+	var (
+		modelLoader *model.ModelLoader
+		modelPath   string
+		mockModel   *model.Model
+	)
 
-const llama3 = `<|start_header_id|>{{if eq .RoleName "assistant"}}assistant{{else if eq .RoleName "system"}}system{{else if eq .RoleName "tool"}}tool{{else if eq .RoleName "user"}}user{{end}}<|end_header_id|>
-
-{{ if .FunctionCall -}}
-Function call:
-{{ else if eq .RoleName "tool" -}}
-Function response:
-{{ end -}}
-{{ if .Content -}}
-{{.Content -}}
-{{ else if .FunctionCall -}}
-{{ toJson .FunctionCall -}}
-{{ end -}}
-<|eot_id|>`
-
-var llama3TestMatch map[string]map[string]interface{} = map[string]map[string]interface{}{
-	"user": {
-		"template": llama3,
-		"expected": "<|start_header_id|>user<|end_header_id|>\n\nA long time ago in a galaxy far, far away...<|eot_id|>",
-		"data": ChatMessageTemplateData{
-			SystemPrompt: "",
-			Role:         "user",
-			RoleName:     "user",
-			Content:      "A long time ago in a galaxy far, far away...",
-			FunctionCall: nil,
-			FunctionName: "",
-			LastMessage:  false,
-			Function:     false,
-			MessageIndex: 0,
-		},
-	},
-	"assistant": {
-		"template": llama3,
-		"expected": "<|start_header_id|>assistant<|end_header_id|>\n\nA long time ago in a galaxy far, far away...<|eot_id|>",
-		"data": ChatMessageTemplateData{
-			SystemPrompt: "",
-			Role:         "assistant",
-			RoleName:     "assistant",
-			Content:      "A long time ago in a galaxy far, far away...",
-			FunctionCall: nil,
-			FunctionName: "",
-			LastMessage:  false,
-			Function:     false,
-			MessageIndex: 0,
-		},
-	},
-	"function_call": {
-		"template": llama3,
-		"expected": "<|start_header_id|>assistant<|end_header_id|>\n\nFunction call:\n{\"function\":\"test\"}<|eot_id|>",
-		"data": ChatMessageTemplateData{
-			SystemPrompt: "",
-			Role:         "assistant",
-			RoleName:     "assistant",
-			Content:      "",
-			FunctionCall: map[string]string{"function": "test"},
-			FunctionName: "",
-			LastMessage:  false,
-			Function:     false,
-			MessageIndex: 0,
-		},
-	},
-	"function_response": {
-		"template": llama3,
-		"expected": "<|start_header_id|>tool<|end_header_id|>\n\nFunction response:\nResponse from tool<|eot_id|>",
-		"data": ChatMessageTemplateData{
-			SystemPrompt: "",
-			Role:         "tool",
-			RoleName:     "tool",
-			Content:      "Response from tool",
-			FunctionCall: nil,
-			FunctionName: "",
-			LastMessage:  false,
-			Function:     false,
-			MessageIndex: 0,
-		},
-	},
-}
-
-var chatMLTestMatch map[string]map[string]interface{} = map[string]map[string]interface{}{
-	"user": {
-		"template": chatML,
-		"expected": "<|im_start|>user\nA long time ago in a galaxy far, far away...<|im_end|>",
-		"data": ChatMessageTemplateData{
-			SystemPrompt: "",
-			Role:         "user",
-			RoleName:     "user",
-			Content:      "A long time ago in a galaxy far, far away...",
-			FunctionCall: nil,
-			FunctionName: "",
-			LastMessage:  false,
-			Function:     false,
-			MessageIndex: 0,
-		},
-	},
-	"assistant": {
-		"template": chatML,
-		"expected": "<|im_start|>assistant\nA long time ago in a galaxy far, far away...<|im_end|>",
-		"data": ChatMessageTemplateData{
-			SystemPrompt: "",
-			Role:         "assistant",
-			RoleName:     "assistant",
-			Content:      "A long time ago in a galaxy far, far away...",
-			FunctionCall: nil,
-			FunctionName: "",
-			LastMessage:  false,
-			Function:     false,
-			MessageIndex: 0,
-		},
-	},
-	"function_call": {
-		"template": chatML,
-		"expected": "<|im_start|>assistant\n<tool_call>\n{\"function\":\"test\"}\n</tool_call><|im_end|>",
-		"data": ChatMessageTemplateData{
-			SystemPrompt: "",
-			Role:         "assistant",
-			RoleName:     "assistant",
-			Content:      "",
-			FunctionCall: map[string]string{"function": "test"},
-			FunctionName: "",
-			LastMessage:  false,
-			Function:     false,
-			MessageIndex: 0,
-		},
-	},
-	"function_response": {
-		"template": chatML,
-		"expected": "<|im_start|>tool\n<tool_response>\nResponse from tool\n</tool_response><|im_end|>",
-		"data": ChatMessageTemplateData{
-			SystemPrompt: "",
-			Role:         "tool",
-			RoleName:     "tool",
-			Content:      "Response from tool",
-			FunctionCall: nil,
-			FunctionName: "",
-			LastMessage:  false,
-			Function:     false,
-			MessageIndex: 0,
-		},
-	},
-}
-
-var _ = Describe("Templates", func() {
-	Context("chat message ChatML", func() {
-		var modelLoader *ModelLoader
-		BeforeEach(func() {
-			modelLoader = NewModelLoader("")
-		})
-		for key := range chatMLTestMatch {
-			foo := chatMLTestMatch[key]
-			It("renders correctly `"+key+"`", func() {
-				templated, err := modelLoader.EvaluateTemplateForChatMessage(foo["template"].(string), foo["data"].(ChatMessageTemplateData))
-				Expect(err).ToNot(HaveOccurred())
-				Expect(templated).To(Equal(foo["expected"]), templated)
-			})
-		}
+	BeforeEach(func() {
+		// Setup the model loader with a test directory
+		modelPath = "/tmp/test_model_path"
+		os.Mkdir(modelPath, 0755)
+		modelLoader = model.NewModelLoader(modelPath)
 	})
-	Context("chat message llama3", func() {
-		var modelLoader *ModelLoader
-		BeforeEach(func() {
-			modelLoader = NewModelLoader("")
+
+	AfterEach(func() {
+		// Cleanup test directory
+		os.RemoveAll(modelPath)
+	})
+
+	Context("NewModelLoader", func() {
+		It("should create a new ModelLoader with an empty model map", func() {
+			Expect(modelLoader).ToNot(BeNil())
+			Expect(modelLoader.ModelPath).To(Equal(modelPath))
+			Expect(modelLoader.ListModels()).To(BeEmpty())
+		})
+	})
+
+	Context("ExistsInModelPath", func() {
+		It("should return true if a file exists in the model path", func() {
+			testFile := filepath.Join(modelPath, "test.model")
+			os.Create(testFile)
+			Expect(modelLoader.ExistsInModelPath("test.model")).To(BeTrue())
+		})
+
+		It("should return false if a file does not exist in the model path", func() {
+			Expect(modelLoader.ExistsInModelPath("nonexistent.model")).To(BeFalse())
+		})
+	})
+
+	Context("ListFilesInModelPath", func() {
+		It("should list all valid model files in the model path", func() {
+			os.Create(filepath.Join(modelPath, "test.model"))
+			os.Create(filepath.Join(modelPath, "README.md"))
+
+			files, err := modelLoader.ListFilesInModelPath()
+			Expect(err).To(BeNil())
+			Expect(files).To(ContainElement("test.model"))
+			Expect(files).ToNot(ContainElement("README.md"))
+		})
+	})
+
+	Context("LoadModel", func() {
+		It("should load a model and keep it in memory", func() {
+			mockModel = model.NewModel("test.model")
+
+			mockLoader := func(modelName, modelFile string) (*model.Model, error) {
+				return mockModel, nil
+			}
+
+			model, err := modelLoader.LoadModel("test.model", mockLoader)
+			Expect(err).To(BeNil())
+			Expect(model).To(Equal(mockModel))
+			Expect(modelLoader.CheckIsLoaded("test.model")).To(Equal(mockModel))
+		})
+
+		It("should return an error if loading the model fails", func() {
+			mockLoader := func(modelName, modelFile string) (*model.Model, error) {
+				return nil, errors.New("failed to load model")
+			}
+
+			model, err := modelLoader.LoadModel("test.model", mockLoader)
+			Expect(err).To(HaveOccurred())
+			Expect(model).To(BeNil())
+		})
+	})
+
+	Context("ShutdownModel", func() {
+		It("should shutdown a loaded model", func() {
+			mockModel = model.NewModel("test.model")
+
+			mockLoader := func(modelName, modelFile string) (*model.Model, error) {
+				return mockModel, nil
+			}
+
+			_, err := modelLoader.LoadModel("test.model", mockLoader)
+			Expect(err).To(BeNil())
+
+			err = modelLoader.ShutdownModel("test.model")
+			Expect(err).To(BeNil())
+			Expect(modelLoader.CheckIsLoaded("test.model")).To(BeNil())
 		})
-		for key := range llama3TestMatch {
-			foo := llama3TestMatch[key]
-			It("renders correctly `"+key+"`", func() {
-				templated, err := modelLoader.EvaluateTemplateForChatMessage(foo["template"].(string), foo["data"].(ChatMessageTemplateData))
-				Expect(err).ToNot(HaveOccurred())
-				Expect(templated).To(Equal(foo["expected"]), templated)
-			})
-		}
 	})
 })
diff --git a/pkg/model/model.go b/pkg/model/model.go
new file mode 100644
index 00000000..1927dc0c
--- /dev/null
+++ b/pkg/model/model.go
@@ -0,0 +1,28 @@
+package model
+
+import grpc "github.com/mudler/LocalAI/pkg/grpc"
+
+type Model struct {
+	address string
+	client  grpc.Backend
+}
+
+func NewModel(address string) *Model {
+	return &Model{
+		address: address,
+	}
+}
+
+func (m *Model) GRPC(parallel bool, wd *WatchDog) grpc.Backend {
+	if m.client != nil {
+		return m.client
+	}
+
+	enableWD := false
+	if wd != nil {
+		enableWD = true
+	}
+
+	m.client = grpc.NewClient(m.address, parallel, wd, enableWD)
+	return m.client
+}
diff --git a/pkg/model/process.go b/pkg/model/process.go
index 7b7ecb97..50afbb1c 100644
--- a/pkg/model/process.go
+++ b/pkg/model/process.go
@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"os"
 	"os/signal"
+	"path/filepath"
 	"strconv"
 	"strings"
 	"syscall"
@@ -17,22 +18,23 @@ import (
 
 func (ml *ModelLoader) StopAllExcept(s string) error {
 	return ml.StopGRPC(func(id string, p *process.Process) bool {
-		if id != s {
-			for ml.models[id].GRPC(false, ml.wd).IsBusy() {
-				log.Debug().Msgf("%s busy. Waiting.", id)
-				time.Sleep(2 * time.Second)
-			}
-			log.Debug().Msgf("[single-backend] Stopping %s", id)
-			return true
+		if id == s {
+			return false
 		}
-		return false
+
+		for ml.models[id].GRPC(false, ml.wd).IsBusy() {
+			log.Debug().Msgf("%s busy. Waiting.", id)
+			time.Sleep(2 * time.Second)
+		}
+		log.Debug().Msgf("[single-backend] Stopping %s", id)
+		return true
 	})
 }
 
 func (ml *ModelLoader) deleteProcess(s string) error {
 	if _, exists := ml.grpcProcesses[s]; exists {
 		if err := ml.grpcProcesses[s].Stop(); err != nil {
-			return err
+			log.Error().Err(err).Msgf("(deleteProcess) error while deleting grpc process %s", s)
 		}
 	}
 	delete(ml.grpcProcesses, s)
@@ -79,11 +81,17 @@ func (ml *ModelLoader) startProcess(grpcProcess, id string, serverAddress string
 
 	log.Debug().Msgf("GRPC Service for %s will be running at: '%s'", id, serverAddress)
 
+	workDir, err := filepath.Abs(filepath.Dir(grpcProcess))
+	if err != nil {
+		return err
+	}
+
 	grpcControlProcess := process.New(
 		process.WithTemporaryStateDir(),
-		process.WithName(grpcProcess),
+		process.WithName(filepath.Base(grpcProcess)),
 		process.WithArgs(append(args, []string{"--addr", serverAddress}...)...),
 		process.WithEnvironment(os.Environ()...),
+		process.WithWorkDir(workDir),
 	)
 
 	if ml.wd != nil {
diff --git a/pkg/model/template.go b/pkg/model/template.go
new file mode 100644
index 00000000..3dc850cf
--- /dev/null
+++ b/pkg/model/template.go
@@ -0,0 +1,52 @@
+package model
+
+import (
+	"fmt"
+
+	"github.com/mudler/LocalAI/pkg/functions"
+	"github.com/mudler/LocalAI/pkg/templates"
+)
+
+// Rather than pass an interface{} to the prompt template:
+// These are the definitions of all possible variables LocalAI will currently populate for use in a prompt template file
+// Please note: Not all of these are populated on every endpoint - your template should either be tested for each endpoint you map it to, or tolerant of zero values.
+type PromptTemplateData struct {
+	SystemPrompt         string
+	SuppressSystemPrompt bool // used by chat specifically to indicate that SystemPrompt above should be _ignored_
+	Input                string
+	Instruction          string
+	Functions            []functions.Function
+	MessageIndex         int
+}
+
+type ChatMessageTemplateData struct {
+	SystemPrompt string
+	Role         string
+	RoleName     string
+	FunctionName string
+	Content      string
+	MessageIndex int
+	Function     bool
+	FunctionCall interface{}
+	LastMessage  bool
+}
+
+const (
+	ChatPromptTemplate templates.TemplateType = iota
+	ChatMessageTemplate
+	CompletionPromptTemplate
+	EditPromptTemplate
+	FunctionsPromptTemplate
+)
+
+func (ml *ModelLoader) EvaluateTemplateForPrompt(templateType templates.TemplateType, templateName string, in PromptTemplateData) (string, error) {
+	// TODO: should this check be improved?
+	if templateType == ChatMessageTemplate {
+		return "", fmt.Errorf("invalid templateType: ChatMessage")
+	}
+	return ml.templates.EvaluateTemplate(templateType, templateName, in)
+}
+
+func (ml *ModelLoader) EvaluateTemplateForChatMessage(templateName string, messageData ChatMessageTemplateData) (string, error) {
+	return ml.templates.EvaluateTemplate(ChatMessageTemplate, templateName, messageData)
+}
diff --git a/pkg/model/template_test.go b/pkg/model/template_test.go
new file mode 100644
index 00000000..1142ed0c
--- /dev/null
+++ b/pkg/model/template_test.go
@@ -0,0 +1,197 @@
+package model_test
+
+import (
+	. "github.com/mudler/LocalAI/pkg/model"
+
+	. "github.com/onsi/ginkgo/v2"
+	. "github.com/onsi/gomega"
+)
+
+const chatML = `<|im_start|>{{if eq .RoleName "assistant"}}assistant{{else if eq .RoleName "system"}}system{{else if eq .RoleName "tool"}}tool{{else if eq .RoleName "user"}}user{{end}}
+{{- if .FunctionCall }}
+<tool_call>
+{{- else if eq .RoleName "tool" }}
+<tool_response>
+{{- end }}
+{{- if .Content}}
+{{.Content }}
+{{- end }}
+{{- if .FunctionCall}}
+{{toJson .FunctionCall}}
+{{- end }}
+{{- if .FunctionCall }}
+</tool_call>
+{{- else if eq .RoleName "tool" }}
+</tool_response>
+{{- end }}<|im_end|>`
+
+const llama3 = `<|start_header_id|>{{if eq .RoleName "assistant"}}assistant{{else if eq .RoleName "system"}}system{{else if eq .RoleName "tool"}}tool{{else if eq .RoleName "user"}}user{{end}}<|end_header_id|>
+
+{{ if .FunctionCall -}}
+Function call:
+{{ else if eq .RoleName "tool" -}}
+Function response:
+{{ end -}}
+{{ if .Content -}}
+{{.Content -}}
+{{ else if .FunctionCall -}}
+{{ toJson .FunctionCall -}}
+{{ end -}}
+<|eot_id|>`
+
+var llama3TestMatch map[string]map[string]interface{} = map[string]map[string]interface{}{
+	"user": {
+		"template": llama3,
+		"expected": "<|start_header_id|>user<|end_header_id|>\n\nA long time ago in a galaxy far, far away...<|eot_id|>",
+		"data": ChatMessageTemplateData{
+			SystemPrompt: "",
+			Role:         "user",
+			RoleName:     "user",
+			Content:      "A long time ago in a galaxy far, far away...",
+			FunctionCall: nil,
+			FunctionName: "",
+			LastMessage:  false,
+			Function:     false,
+			MessageIndex: 0,
+		},
+	},
+	"assistant": {
+		"template": llama3,
+		"expected": "<|start_header_id|>assistant<|end_header_id|>\n\nA long time ago in a galaxy far, far away...<|eot_id|>",
+		"data": ChatMessageTemplateData{
+			SystemPrompt: "",
+			Role:         "assistant",
+			RoleName:     "assistant",
+			Content:      "A long time ago in a galaxy far, far away...",
+			FunctionCall: nil,
+			FunctionName: "",
+			LastMessage:  false,
+			Function:     false,
+			MessageIndex: 0,
+		},
+	},
+	"function_call": {
+		"template": llama3,
+		"expected": "<|start_header_id|>assistant<|end_header_id|>\n\nFunction call:\n{\"function\":\"test\"}<|eot_id|>",
+		"data": ChatMessageTemplateData{
+			SystemPrompt: "",
+			Role:         "assistant",
+			RoleName:     "assistant",
+			Content:      "",
+			FunctionCall: map[string]string{"function": "test"},
+			FunctionName: "",
+			LastMessage:  false,
+			Function:     false,
+			MessageIndex: 0,
+		},
+	},
+	"function_response": {
+		"template": llama3,
+		"expected": "<|start_header_id|>tool<|end_header_id|>\n\nFunction response:\nResponse from tool<|eot_id|>",
+		"data": ChatMessageTemplateData{
+			SystemPrompt: "",
+			Role:         "tool",
+			RoleName:     "tool",
+			Content:      "Response from tool",
+			FunctionCall: nil,
+			FunctionName: "",
+			LastMessage:  false,
+			Function:     false,
+			MessageIndex: 0,
+		},
+	},
+}
+
+var chatMLTestMatch map[string]map[string]interface{} = map[string]map[string]interface{}{
+	"user": {
+		"template": chatML,
+		"expected": "<|im_start|>user\nA long time ago in a galaxy far, far away...<|im_end|>",
+		"data": ChatMessageTemplateData{
+			SystemPrompt: "",
+			Role:         "user",
+			RoleName:     "user",
+			Content:      "A long time ago in a galaxy far, far away...",
+			FunctionCall: nil,
+			FunctionName: "",
+			LastMessage:  false,
+			Function:     false,
+			MessageIndex: 0,
+		},
+	},
+	"assistant": {
+		"template": chatML,
+		"expected": "<|im_start|>assistant\nA long time ago in a galaxy far, far away...<|im_end|>",
+		"data": ChatMessageTemplateData{
+			SystemPrompt: "",
+			Role:         "assistant",
+			RoleName:     "assistant",
+			Content:      "A long time ago in a galaxy far, far away...",
+			FunctionCall: nil,
+			FunctionName: "",
+			LastMessage:  false,
+			Function:     false,
+			MessageIndex: 0,
+		},
+	},
+	"function_call": {
+		"template": chatML,
+		"expected": "<|im_start|>assistant\n<tool_call>\n{\"function\":\"test\"}\n</tool_call><|im_end|>",
+		"data": ChatMessageTemplateData{
+			SystemPrompt: "",
+			Role:         "assistant",
+			RoleName:     "assistant",
+			Content:      "",
+			FunctionCall: map[string]string{"function": "test"},
+			FunctionName: "",
+			LastMessage:  false,
+			Function:     false,
+			MessageIndex: 0,
+		},
+	},
+	"function_response": {
+		"template": chatML,
+		"expected": "<|im_start|>tool\n<tool_response>\nResponse from tool\n</tool_response><|im_end|>",
+		"data": ChatMessageTemplateData{
+			SystemPrompt: "",
+			Role:         "tool",
+			RoleName:     "tool",
+			Content:      "Response from tool",
+			FunctionCall: nil,
+			FunctionName: "",
+			LastMessage:  false,
+			Function:     false,
+			MessageIndex: 0,
+		},
+	},
+}
+
+var _ = Describe("Templates", func() {
+	Context("chat message ChatML", func() {
+		var modelLoader *ModelLoader
+		BeforeEach(func() {
+			modelLoader = NewModelLoader("")
+		})
+		for key := range chatMLTestMatch {
+			foo := chatMLTestMatch[key]
+			It("renders correctly `"+key+"`", func() {
+				templated, err := modelLoader.EvaluateTemplateForChatMessage(foo["template"].(string), foo["data"].(ChatMessageTemplateData))
+				Expect(err).ToNot(HaveOccurred())
+				Expect(templated).To(Equal(foo["expected"]), templated)
+			})
+		}
+	})
+	Context("chat message llama3", func() {
+		var modelLoader *ModelLoader
+		BeforeEach(func() {
+			modelLoader = NewModelLoader("")
+		})
+		for key := range llama3TestMatch {
+			foo := llama3TestMatch[key]
+			It("renders correctly `"+key+"`", func() {
+				templated, err := modelLoader.EvaluateTemplateForChatMessage(foo["template"].(string), foo["data"].(ChatMessageTemplateData))
+				Expect(err).ToNot(HaveOccurred())
+				Expect(templated).To(Equal(foo["expected"]), templated)
+			})
+		}
+	})
+})
diff --git a/pkg/model/watchdog.go b/pkg/model/watchdog.go
index b5381832..5702dda5 100644
--- a/pkg/model/watchdog.go
+++ b/pkg/model/watchdog.go
@@ -8,6 +8,7 @@ import (
 	"github.com/rs/zerolog/log"
 )
 
+// WatchDog tracks all the requests from GRPC clients.
 // All GRPC Clients created by ModelLoader should have an associated injected
 // watchdog that will keep track of the state of each backend (busy or not)
 // and for how much time it has been busy.
@@ -15,7 +16,6 @@ import (
 // force a reload of the model
 // The watchdog runs as a separate go routine,
 // and the GRPC client talks to it via a channel to send status updates
-
 type WatchDog struct {
 	sync.Mutex
 	timetable            map[string]time.Time
diff --git a/pkg/utils/ffmpeg.go b/pkg/utils/ffmpeg.go
new file mode 100644
index 00000000..16656d8e
--- /dev/null
+++ b/pkg/utils/ffmpeg.go
@@ -0,0 +1,25 @@
+package utils
+
+import (
+	"fmt"
+	"os"
+	"os/exec"
+)
+
+func ffmpegCommand(args []string) (string, error) {
+	cmd := exec.Command("ffmpeg", args...) // Constrain this to ffmpeg to permit security scanner to see that the command is safe.
+	cmd.Env = os.Environ()
+	out, err := cmd.CombinedOutput()
+	return string(out), err
+}
+
+// AudioToWav converts audio to wav for transcribe.
+// TODO: use https://github.com/mccoyst/ogg?
+func AudioToWav(src, dst string) error {
+	commandArgs := []string{"-i", src, "-format", "s16le", "-ar", "16000", "-ac", "1", "-acodec", "pcm_s16le", dst}
+	out, err := ffmpegCommand(commandArgs)
+	if err != nil {
+		return fmt.Errorf("error: %w out: %s", err, out)
+	}
+	return nil
+}
diff --git a/pkg/utils/path.go b/pkg/utils/path.go
index c1d3e86d..1ae11d12 100644
--- a/pkg/utils/path.go
+++ b/pkg/utils/path.go
@@ -38,3 +38,19 @@ func SanitizeFileName(fileName string) string {
 	safeName := strings.ReplaceAll(baseName, "..", "")
 	return safeName
 }
+
+func GenerateUniqueFileName(dir, baseName, ext string) string {
+	counter := 1
+	fileName := baseName + ext
+
+	for {
+		filePath := filepath.Join(dir, fileName)
+		_, err := os.Stat(filePath)
+		if os.IsNotExist(err) {
+			return fileName
+		}
+
+		counter++
+		fileName = fmt.Sprintf("%s_%d%s", baseName, counter, ext)
+	}
+}
diff --git a/swagger/docs.go b/swagger/docs.go
index 44ae10ad..44da7cf2 100644
--- a/swagger/docs.go
+++ b/swagger/docs.go
@@ -266,6 +266,19 @@ const docTemplate = `{
                 }
             }
         },
+        "/system": {
+            "get": {
+                "summary": "Show the LocalAI instance information",
+                "responses": {
+                    "200": {
+                        "description": "Response",
+                        "schema": {
+                            "$ref": "#/definitions/schema.SystemInformationResponse"
+                        }
+                    }
+                }
+            }
+        },
         "/tts": {
             "post": {
                 "consumes": [
@@ -656,6 +669,30 @@ const docTemplate = `{
                 }
             }
         },
+        "/v1/sound-generation": {
+            "post": {
+                "summary": "Generates audio from the input text.",
+                "parameters": [
+                    {
+                        "description": "query params",
+                        "name": "request",
+                        "in": "body",
+                        "required": true,
+                        "schema": {
+                            "$ref": "#/definitions/schema.ElevenLabsSoundGenerationRequest"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Response",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
         "/v1/text-to-speech/{voice-id}": {
             "post": {
                 "summary": "Generates audio from the input text.",
@@ -1161,6 +1198,26 @@ const docTemplate = `{
                 }
             }
         },
+        "schema.ElevenLabsSoundGenerationRequest": {
+            "type": "object",
+            "properties": {
+                "do_sample": {
+                    "type": "boolean"
+                },
+                "duration_seconds": {
+                    "type": "number"
+                },
+                "model_id": {
+                    "type": "string"
+                },
+                "prompt_influence": {
+                    "type": "number"
+                },
+                "text": {
+                    "type": "string"
+                }
+            }
+        },
         "schema.File": {
             "type": "object",
             "properties": {
@@ -1605,6 +1662,17 @@ const docTemplate = `{
                 }
             }
         },
+        "schema.SystemInformationResponse": {
+            "type": "object",
+            "properties": {
+                "backends": {
+                    "type": "array",
+                    "items": {
+                        "type": "string"
+                    }
+                }
+            }
+        },
         "schema.TTSRequest": {
             "description": "TTS request body",
             "type": "object",
diff --git a/swagger/swagger.json b/swagger/swagger.json
index 6edfebbd..eaddf451 100644
--- a/swagger/swagger.json
+++ b/swagger/swagger.json
@@ -259,6 +259,19 @@
                 }
             }
         },
+        "/system": {
+            "get": {
+                "summary": "Show the LocalAI instance information",
+                "responses": {
+                    "200": {
+                        "description": "Response",
+                        "schema": {
+                            "$ref": "#/definitions/schema.SystemInformationResponse"
+                        }
+                    }
+                }
+            }
+        },
         "/tts": {
             "post": {
                 "consumes": [
@@ -649,6 +662,30 @@
                 }
             }
         },
+        "/v1/sound-generation": {
+            "post": {
+                "summary": "Generates audio from the input text.",
+                "parameters": [
+                    {
+                        "description": "query params",
+                        "name": "request",
+                        "in": "body",
+                        "required": true,
+                        "schema": {
+                            "$ref": "#/definitions/schema.ElevenLabsSoundGenerationRequest"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Response",
+                        "schema": {
+                            "type": "string"
+                        }
+                    }
+                }
+            }
+        },
         "/v1/text-to-speech/{voice-id}": {
             "post": {
                 "summary": "Generates audio from the input text.",
@@ -1154,6 +1191,26 @@
                 }
             }
         },
+        "schema.ElevenLabsSoundGenerationRequest": {
+            "type": "object",
+            "properties": {
+                "do_sample": {
+                    "type": "boolean"
+                },
+                "duration_seconds": {
+                    "type": "number"
+                },
+                "model_id": {
+                    "type": "string"
+                },
+                "prompt_influence": {
+                    "type": "number"
+                },
+                "text": {
+                    "type": "string"
+                }
+            }
+        },
         "schema.File": {
             "type": "object",
             "properties": {
@@ -1598,6 +1655,17 @@
                 }
             }
         },
+        "schema.SystemInformationResponse": {
+            "type": "object",
+            "properties": {
+                "backends": {
+                    "type": "array",
+                    "items": {
+                        "type": "string"
+                    }
+                }
+            }
+        },
         "schema.TTSRequest": {
             "description": "TTS request body",
             "type": "object",
diff --git a/swagger/swagger.yaml b/swagger/swagger.yaml
index c953b0af..c98e0ef4 100644
--- a/swagger/swagger.yaml
+++ b/swagger/swagger.yaml
@@ -322,6 +322,19 @@ definitions:
       object:
         type: string
     type: object
+  schema.ElevenLabsSoundGenerationRequest:
+    properties:
+      do_sample:
+        type: boolean
+      duration_seconds:
+        type: number
+      model_id:
+        type: string
+      prompt_influence:
+        type: number
+      text:
+        type: string
+    type: object
   schema.File:
     properties:
       bytes:
@@ -625,6 +638,13 @@ definitions:
           $ref: '#/definitions/p2p.NodeData'
         type: array
     type: object
+  schema.SystemInformationResponse:
+    properties:
+      backends:
+        items:
+          type: string
+        type: array
+    type: object
   schema.TTSRequest:
     description: TTS request body
     properties:
@@ -819,6 +839,14 @@ paths:
           schema:
             $ref: '#/definitions/gallery.GalleryOpStatus'
       summary: Returns the job status
+  /system:
+    get:
+      responses:
+        "200":
+          description: Response
+          schema:
+            $ref: '#/definitions/schema.SystemInformationResponse'
+      summary: Show the LocalAI instance information
   /tts:
     post:
       consumes:
@@ -1066,6 +1094,21 @@ paths:
           schema:
             $ref: '#/definitions/schema.JINARerankResponse'
       summary: Reranks a list of phrases by relevance to a given text query.
+  /v1/sound-generation:
+    post:
+      parameters:
+      - description: query params
+        in: body
+        name: request
+        required: true
+        schema:
+          $ref: '#/definitions/schema.ElevenLabsSoundGenerationRequest'
+      responses:
+        "200":
+          description: Response
+          schema:
+            type: string
+      summary: Generates audio from the input text.
   /v1/text-to-speech/{voice-id}:
     post:
       parameters: