wtf is llamaforge?

LlamaForge is a graphical dashboard that sits on top of llama.cpp, the popular tool for running large language models locally on your own computer. Instead of hand editing configuration files or typing long command lines to start a model server, you manage everything from a browser based control panel: building llama.cpp, keeping it updated with the latest version from its source, finding models that fit your hardware, adjusting every setting for a model, and running it. The Models tab lists every model on your machine along with live graphics card memory and temperature readings. You can expand a model to adjust around 220 different llama.cpp settings, grouped and searchable, see details pulled from the model file itself such as its architecture and quantization, save a group of settings as a named preset, compare two or three models side by side, and copy ready to use code snippets for calling the model from your own programs. The Stats tab tracks how much each model has been used, how fast it generates text, and how long it has been loaded. The Discover tab searches Hugging Face for compatible models and rates each one against your available graphics memory, showing whether it will fit comfortably, fit tightly, or need to run partly on the CPU, before you download it. The Build and Update tab checks how far behind the latest llama.cpp version you are and rebuilds it with settings tuned for your specific CPU or graphics card. A Setup tab checks for required tools like Git and CMake and can install missing ones with your permission. Beyond llama.cpp, LlamaForge can also run vLLM, a second engine used for full precision models in different formats, with the two engines sharing the same model list and interface. On Windows, this second engine runs inside a Linux compatibility layer called WSL2. This project is aimed at people who want the speed and fine grained control of llama.cpp but do not want to memorize command line flags or manage build steps by hand. It targets Windows with an NVIDIA graphics card as the primary setup, with Linux and macOS supported as an early preview. It is an independent project, not affiliated with llama.cpp itself, and is released under the MIT license.