LlamaChat

LlamaChat is a free, open source macOS app for chatting with LLaMA family language models that run locally on the user's own Mac. It supports LLaMA, Alpaca and GPT4All models, and lists Vicuna as coming soon. Because inference happens on the device, conversations do not need to be sent to a cloud service. The app is aimed at developers and enthusiasts who want to experiment with open models without writing code. It can import raw published PyTorch model checkpoints directly, or models that are already converted to the .ggml format, so conversion does not have to be done by hand. It is built on the open source libraries llama.cpp and llama.swift. It is built for both Intel processors and Apple Silicon, and version 1.2.0 requires macOS 13. It can be downloaded from the site or installed with Homebrew using the cask named llamachat. LlamaChat does not ship with any model files, proprietary or otherwise. Users are responsible for obtaining the models and complying with the terms set by their providers. The site states that the app is an independent project, not affiliated with or endorsed by Meta, Stanford or Nomic AI. The site states that LlamaChat is 100 percent free and fully open source, and that it always will be. There are no paid plans. The source code is on GitHub and the project accepts pull requests.