- **Revamped Main Settings Structure**: The primary settings interface has been reorganized to improve usability and maintainability, providing streamlined access to essential configuration options.
- **Overhauled Services System Architecture**: The services framework has been comprehensively restructured to improve modularity and scalability, facilitating better integration and management of various service types.
- **Enhanced Service Parameter Flexibility**: Service parameters have been redesigned to support dynamic customization, accommodating a broader range of services with configurable options for precise control over functionality.
- **Expanded Function Call Capabilities**: Function calls now support both static and dynamic parameters. Static parameters ensure consistent predefined values, while dynamic parameters provide adaptability, enhancing control and versatility in execution.
- **System-Wide Function Calling**: Function calling capabilities have been fully implemented across the system for consistent operation.
- **Full Integration with Lightrag**: Seamless integration with Lightrag has been achieved, enhancing system functionality.
- **Enhanced User Interface**: The UI has been refined for improved usability and aesthetics.
- **Reworked Context Management System**: The context management framework has been optimized for better performance and reliability.
- **Long-Term Memory Reintroduced**: Long-term memory functionality has been restored through a dedicated function call mechanism.
- **[Video Tutorial](https://youtu.be/0ft6PyOvSqI?si=3bFtOzQ-J2Y53JaY)**: A detailed guide on function calls in LoLLMs is available for user reference.
- **Discontinued Raw HTML Rendering**: Messages no longer render raw HTML, bolstering security against code injection vulnerabilities. This update aligns with LoLLMs’ compatibility with Docker environments.
This release demonstrates that advanced thinking capabilities can be achieved with traditional LLMs through effective prompting and visualization, without reliance on specialized models such as O1 or R1.