MLXHub: Local AI & LLM Server

An AI that lives on your iPhone and iPad, even without internet. Talk to it and hear it answer out loud, create images right on your device, and turn it into a server for your local network. No message counter.

Talk to it. It talks back.
Voice mode transcribes what you say on your device, in the languages you have set up, answers with a natural voice and lets you cut in whenever you want.

An assistant that can help you act
Ask about your calendar, reminders, Health data, photos or music library with supported tools; you choose which ones to authorize. With MLXHub Plus, create events and reminders, log Health data and save reusable instructions as Skills. Tools and results depend on the model you choose.

Memory you control
Keep what is worth remembering for future chats, and review, pin or delete it in Settings. With Plus, compatible models can also save or remove memories on their own.

Create images on your device
Download a compatible image model and generate images without uploading anything to the cloud, from the Image tab or by asking in a chat. Pause an image and resume it later. Free up to 512 px; up to 1280 px with Plus.

Bigger open models
Download open-weight language and vision models. Supported MoE models load their experts from storage when needed, which is how a 35B MoE runs on iPhone 17 Pro Max and on supported iPads. Capacity and performance depend on your device, model and context.

Text, photos and documents
Write, attach a photo or a document and ask a compatible model for help, with the personality and instructions you set. On devices with Apple Intelligence you can also use Apple's on-device model.

Measure your device
The benchmark scores the speed of your iPhone or iPad with real models. Publish your result to the leaderboard and compare it with other devices.

Your devices, working together
Connect compatible devices on your local network to run models across them. Joining a mesh is free; hosting one requires MLXHub Plus.

A local server for your apps
Serve chat and embeddings to apps and scripts on your local network through a standard chat completions API (/v1/chat/completions, /v1/embeddings), with authentication and local discovery. It includes a web chat portal for any browser on your Wi-Fi and a voice center to talk to the server's model. Unlock the server with a one-time purchase or with MLXHub Plus; it runs while the app is in the foreground.

All yours, unmetered
No token limits, no message caps, no monthly quota, and no AI company keeping an eye on what you write. Inference happens inside your device, and your conversations are not uploaded to any cloud AI service: an assistant and a server you can use without limits, under your control.

And if you would rather not subscribe, the one-time Lifetime purchase gives you unlimited inference for life and every feature that lands later, at no extra cost. Because this keeps getting better: with every update, every week with each new open-weight model and every year w