Hexllama
A native desktop interface for discovering, downloading, configuring and serving local large language models through llama.cpp, without hand-editing command-line arguments.
Hexllama is a desktop application that streamlines running local language models with llama.cpp, replacing manual file management and command-line arguments with a unified workspace. Its integrated model hub searches Hugging Face from inside the app, and a download manager can pause, resume or cancel large GGUF downloads and generate an execution template with recommended threads, batch size and context window based on a model's parameters and quantization.
Saved templates can run multiple models at once on different ports, either opening llama.cpp's built-in chat web interface or serving the model silently as an API. Hexllama also manages multiple llama.cpp backend builds, checking the upstream repository for new releases and letting a user download and switch between them, and a visual command editor exposes llama.cpp's server flags as toggles and typed fields instead of memorized arguments.