diff --git a/core/backend/tokenize.go b/core/backend/tokenize.go
new file mode 100644
index 00000000..3c78b17f
--- /dev/null
+++ b/core/backend/tokenize.go
@@ -0,0 +1,50 @@
+package backend
+
+import (
+	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/schema"
+	"github.com/mudler/LocalAI/pkg/grpc"
+	model "github.com/mudler/LocalAI/pkg/model"
+)
+
+func ModelTokenize(s string, loader *model.ModelLoader, backendConfig config.BackendConfig, appConfig *config.ApplicationConfig) (schema.TokenizeResponse, error) {
+
+	modelFile := backendConfig.Model
+
+	grpcOpts := GRPCModelOpts(backendConfig)
+
+	var inferenceModel grpc.Backend
+	var err error
+
+	opts := modelOpts(backendConfig, appConfig, []model.Option{
+		model.WithLoadGRPCLoadModelOpts(grpcOpts),
+		model.WithThreads(uint32(*backendConfig.Threads)),
+		model.WithAssetDir(appConfig.AssetsDestination),
+		model.WithModel(modelFile),
+		model.WithContext(appConfig.Context),
+	})
+
+	if backendConfig.Backend == "" {
+		inferenceModel, err = loader.GreedyLoader(opts...)
+	} else {
+		opts = append(opts, model.WithBackendString(backendConfig.Backend))
+		inferenceModel, err = loader.BackendLoader(opts...)
+	}
+	if err != nil {
+		return schema.TokenizeResponse{}, err
+	}
+
+	predictOptions := gRPCPredictOpts(backendConfig, loader.ModelPath)
+	predictOptions.Prompt = s
+
+	// tokenize the string
+	resp, err := inferenceModel.TokenizeString(appConfig.Context, predictOptions)
+	if err != nil {
+		return schema.TokenizeResponse{}, err
+	}
+
+	return schema.TokenizeResponse{
+		Tokens: resp.Tokens,
+	}, nil
+
+}
diff --git a/core/http/endpoints/localai/tokenize.go b/core/http/endpoints/localai/tokenize.go
new file mode 100644
index 00000000..da110bf8
--- /dev/null
+++ b/core/http/endpoints/localai/tokenize.go
@@ -0,0 +1,58 @@
+package localai
+
+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"
+)
+
+// TokenizeEndpoint exposes a REST API to tokenize the content
+// @Summary Tokenize the input.
+// @Success 200 {object} schema.TokenizeResponse "Response"
+// @Router /v1/tokenize [post]
+func TokenizeEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, appConfig *config.ApplicationConfig) func(c *fiber.Ctx) error {
+	return func(c *fiber.Ctx) error {
+
+		input := new(schema.TokenizeRequest)
+
+		// Get input data from the request body
+		if err := c.BodyParser(input); err != nil {
+			return err
+		}
+
+		modelFile, err := fiberContext.ModelFromContext(c, cl, ml, input.Model, false)
+		if err != nil {
+			modelFile = input.Model
+			log.Warn().Msgf("Model not found in context: %s", input.Model)
+		}
+
+		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 {
+			log.Err(err)
+			modelFile = input.Model
+			log.Warn().Msgf("Model not found in context: %s", input.Model)
+		} else {
+			modelFile = cfg.Model
+		}
+		log.Debug().Msgf("Request for model: %s", modelFile)
+
+		tokenResponse, err := backend.ModelTokenize(input.Content, ml, *cfg, appConfig)
+		if err != nil {
+			return err
+		}
+
+		c.JSON(tokenResponse)
+		return nil
+
+	}
+}
diff --git a/core/http/routes/localai.go b/core/http/routes/localai.go
index 2f65e779..f2f0dfa4 100644
--- a/core/http/routes/localai.go
+++ b/core/http/routes/localai.go
@@ -63,4 +63,7 @@ func RegisterLocalAIRoutes(app *fiber.App,
 
 	app.Get("/system", localai.SystemInformations(ml, appConfig))
 
+	// misc
+	app.Post("/v1/tokenize", localai.TokenizeEndpoint(cl, ml, appConfig))
+
 }
diff --git a/core/schema/tokenize.go b/core/schema/tokenize.go
new file mode 100644
index 00000000..3770cc5a
--- /dev/null
+++ b/core/schema/tokenize.go
@@ -0,0 +1,10 @@
+package schema
+
+type TokenizeRequest struct {
+	Content string `json:"content"`
+	Model   string `json:"model"`
+}
+
+type TokenizeResponse struct {
+	Tokens []int32 `json:"tokens"`
+}