Newelle, the AI-powered assistant for Linux systems, has been updated to version 1.2, bringing with it a host of new features and improvements. ## Key Features * **llama.cpp Integration:** Added support for llama.cpp, with options to recompile with any backend. * **Model Library:** Implemented a new model library for ollama and llama.cpp. * **Hybrid Search:** Improved hybrid search, with a focus on optimizing document reading. * **Command Execution:** Added a tool for executing commands directly from the assistant. * **Tool Groups:** Introduced the ability to create and manage tool groups. * **MCP Server:** Optimized the addition of MCP servers, with STDIO support for non-Flatpak environments. * **Semantic Memory:** Implemented semantic memory management. * **Chat Import/Export:** Added the ability to import and export chats. * **RAG Index:** Ability to add custom folders to the RAG index. * **Message Information:** Improved the message information menu, displaying token count and generation speed. Newelle 1.2 is available for download via [FlatHub](https://flathub.org/en/apps/io.github.qwersyk.Newelle).