Swama

About

Swama is a local AI runtime for Apple Silicon Macs, written in Swift on Apple's MLX. It runs language, vision, embedding, speech-recognition, and text-to-speech models on your Mac and serves them through an OpenAI-compatible API, a command-line tool, and a menu bar app. It requires an Apple Silicon Mac running macOS 15.4 or later.

The API supports chat completions (with streaming, tool calling, and image input), a stateless subset of Responses, embeddings, audio transcription, and text-to-speech (experimental). The /v1/decisions endpoint follows the OpenAI Decisions API and scores predicates, choices, and ratings without generating text. Common models can be run by short alias, such as "swama run qwen3.5", which downloads the model from Hugging Face on first use. Any MLX model on Hugging Face can also be used by its full ID.

Swama can be installed with Homebrew ("brew install swama") or by downloading Swama.dmg from Releases. The menu bar app runs the server in the background, installs the swama command, and sets the context limit. The server listens on port 28100 by default and binds to 0.0.0.0, so it is reachable from other machines on your network unless you pass --host 127.0.0.1.