Happy slice of mcp pizzaMCP.pizza
omlx

omlx

MCP.Pizza Chef: jundot

This app lets you run AI models directly on your Mac's Apple Silicon chip, so your conversations and files never have to leave your computer. It lives in your menu bar, where a guided setup screen walks you through picking a folder for your models, starting the server, and downloading your first one. Once it's running, coding assistants such as Claude Code and GitHub Copilot, along with several other AI apps, can connect to it and use your local models for chat or coding help — no subscription and no internet connection required after the initial download.

Coding
Other

Use This MCP server To

Run AI chat models on my Mac without an internet connection Keep my conversations and files completely private and offline Give my coding assistant a free local AI model to use Chat with an AI model right from my menu bar Download and manage several different AI models in one place Avoid paying a monthly AI subscription by running models myself Check how much memory and speed my Mac is using while chatting

README

oMLX

oMLX

LLM inference, optimized for your Mac
Continuous batching and tiered KV caching, managed directly from your menu bar.

Buy Me A Coffee

License Python 3.11-3.13 Apple Silicon

junkim.dot@gmail.com · https://omlx.ai/me

Install · Quickstart · Features · Models · CLI Configuration · Benchmarks · oMLX.ai

English · 中文 · 한국어 · 日本語

oMLX Admin Dashboard

Every LLM server I tried made me choose between convenience and control. I wanted to pin everyday models in memory, auto-swap heavier ones on demand, set context limits - and manage it all from a menu bar.

oMLX persists KV cache across a hot in-memory tier and cold SSD tier - even when context changes mid-conversation, all past context stays cached and reusable across requests, making local LLMs practical for real coding work with tools like Claude Code. That's why I built it.

Install

macOS App

omlx FAQ

Do I need an account or an API key to use this?
No — oMLX runs entirely on your own Mac, so there's no sign-up and no ongoing subscription. A few specialized models may ask for a free Hugging Face account if they're gated, but the app itself works without one.
Which apps can I connect this to?
You can connect coding assistants such as Claude Code and GitHub Copilot, along with other AI apps that support local AI servers, since oMLX speaks the same language as OpenAI's API.
Can I use this to chat with an AI without an internet connection?
Yes — once you've downloaded a model, you can chat with it completely offline using the built-in chat window, and your conversations never leave your computer.
Can I use this to give my coding assistant a free local AI model?
Yes — after installing the optional add-on, tools like Claude Code can send requests to the models running on your Mac instead of a paid cloud service like OpenAI, Claude, or Gemini.
How hard is it to set up?
The basic app installs like any other Mac app and includes a guided setup screen, but turning on full assistant support takes an extra command-line step, so some comfort with basic computer tasks helps.
What kind of Mac do I need?
You'll need a Mac with an Apple Silicon chip (M1, M2, M3, or M4) running macOS 15 or later.
Will this slow down my Mac?
Running AI models uses a meaningful amount of memory and processing power, especially for larger models, so how smoothly it runs depends on your Mac's specs and which model you pick.
Does it work with Claude, ChatGPT, or Gemini directly?
oMLX runs its own downloaded models rather than connecting to Claude, ChatGPT, or Gemini in the cloud, but assistants such as Claude Code can be pointed at oMLX to use those local models instead of a cloud one.