llama.cpp Manager
A Windows manager for llama.cpp: downloads backends and GGUF models, runs the server and chats.
About
A Windows desktop manager for llama.cpp, built with MyGo and a web frontend that runs in the system WebView2. It detects the GPU and installs a CUDA 12.4, Vulkan or CPU backend, imports GGUF files or downloads them from Hugging Face, and runs llama-server for one model or a whole folder of models. A built-in chat lets you branch off any reply and switch personas.