Apps

The desktop app that made downloading and chatting with local models feel finished.

Discovering, downloading and chatting with local models — including MLX builds — with no terminal at all.

A native SwiftUI chat client that only talks to your own server.

A polished, private front end for models you're already serving from your Mac.

A minimal, offline-first chat app built directly on MLX.

Chatting with small models fully on-device: no server, no account, no network call.

The app that got image generation running smoothly on iPhones first.

Local image generation with SD, SDXL and Flux-class models on Apple silicon.

The app most people mean when they say 'just use Whisper'.

Drag-and-drop audio and video transcription and translation, entirely on-device.

A quantization-focused chat app that sells privacy, not features.

Running quantized open models fully offline across iPhone, iPad and Mac from one app.

A native Mac chat window with a model already inside it.

The shortest path from download to first local answer, with no configuration.

A native client for cloud APIs and local backends in the same window.

Switching between local and cloud models from one native app instead of a browser tab.

The official sample app, which is also a perfectly good chat client.

Trying MLX chat inference natively, and reading the source that makes it work.

A small native app built to show what MLX speed feels like.

Downloading and running MLX community models with no Python environment involved.