This repository has been archived by the owner on Sep 16, 2024. It is now read-only.
All of the frontend and backend code has been rewritten. I've removed support for GGML models, but I've added support for GGUF models, Cuda, OpenCL, MacOS, and Metal. I've also removed unnecessary API endpoints and added more options to the WebUI. The UI is more responsive and dynamic, and it can now render and present markdown code nicely. Overall, the UI is more pleasant and transparent, the API is more intuitive, and the codebase is less confusing.
If you need to access the API documentation, you can find it at here.
Additionally, if you need information about supported models and installation, please visit here.