COMING SOON

Rapid-MLXDesktop.

All the speed, none of the terminal. A menu-bar app for macOS — pick a model, hit start, and every OpenAI-compatible tool on your Mac has a local AI server. Same engine as the CLI.

Watch the repo for the release Meanwhile, the CLI is ready today.
serving

One-click serve

Server status, throughput, and memory live in your menu bar.

Qwen3.5-4B2.4 GB
GPT-OSS 20B11 GB

Model manager

Browse and download models with disk and RAM estimates up front.

Summarize this
On it — locally.

Built-in chat

Talk to any local model instantly — reasoning, vision, tool calls included.

CLI
App

Same engine

CLI and app share models, cache, and config. Nothing duplicated.