MatterWaveHosted language models, one API, every client.

Chat on the web, in the terminal, in VS Code or on the desktop. Build on the same OpenAI-compatible /v1 with the Python and TypeScript SDKs. Run it on a Raspberry Pi when the cloud is not in the loop.

Open the app

Get started

Downloads

One conversation, every client

The web app, the terminal client, the VS Code extension and the desktop app all read and write the same conversations over the same /api routes. Start a thread anywhere and pick it up anywhere else.

An API you already know

/v1 speaks the OpenAI chat-completions wire format. Anything that takes a base_url works against it unchanged; the SDKs add typed responses, streaming iterators, retries and request ids.

Plans that meter in tokens

Every plan comes with a monthly allowance of weighted tokens. The GRT models spend it at 1x, 5x and 20x, so an allowance buys the same dollars of list price whichever model you pick.

Yours to keep

What you send and what comes back is yours. Export everything from Settings, delete it when you want, and nothing you write is used to train models.