Pico MLX Server

Pico MLX Server is a macOS app for people who want to run MLX models locally and use them with an existing AI chat client. It gives MLX Server a graphical interface and exposes local models through an OpenAI-compatible API. The project is open source under the MIT license and requires macOS Sonoma or later.

The app handles the Python environment and MLX Server setup through its interface, reducing the command-line work needed to get a local model server running. Model inference runs on your Mac. Models come from the MLX Community on Hugging Face, including Nous-Hermes-2-Mistral-7B-DPO-4bit-MLX as the default choice.

You can start and stop servers from the menu bar, run multiple servers on separate ports, and view each server's logs in its own window. That makes it useful for trying different models or keeping separate local endpoints available to different clients. A configurable shortcut opens your preferred chat app, with Pico AI Assistant as the default.

Client compatibility depends on the API each app expects: Pico MLX Server uses OpenAI's completions endpoint. It doesn't detect ports that are already occupied, and known interface issues can leave the server status out of date in the server creation window and menu. The official source-build route uses Xcode. The app manages a separate Python/MLX Server environment, using Conda by default; it is not a native Swift inference engine.