MatterWave.
MatterWave is the platform GRT runs on — a chat application with everything you expect: conversations that sync across devices, streaming responses, an OpenAI-compatible API, and apps for web, terminal, VS Code, and desktop.
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 routes. Start a thread anywhere and pick it up anywhere else.
An API you already know.
/v1 uses the OpenAI chat-completions format. Anything that takes a base_url works unchanged; the SDKs add typed responses, streaming, retries, and request ids.
Simple plans.
Sign up, pick a plan, and use every model. Upgrade or cancel whenever you want.
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.
Built for GRT. Open to the rest.
GRT runs here first — it is what the platform was built for. Alongside it, well-known open-weights models serve through the same API, and new ones are added as they are ready.
Our own architecture — weights stored as rotations, tokens as phases. Read the research note →
The list grows as deployments land — adding a model is a deployment plus a routing entry.
Pick your surface.
Web
The full app in the browser at app.matterwave.ai — model picker, streaming, history, settings, export.
Terminal
The TUI client — the same conversations from a shell, with branching, search, and the context meter.
VS Code
The extension brings the chat into the editor — ask about a selection or a file without leaving your workspace.
Desktop
The desktop app for Windows, macOS, and Linux — the web experience, installed, with OS notifications.
Edge
A single on-site box that runs the stack locally and mirrors conversations to the platform — for when the cloud is out of the loop.