RelmioView npm packagePrivate AI relay
Your ChatGPT plan.One clean path to your tools.
Relmio connects a supported ChatGPT/Codex sign-in to the OpenAI-shaped workflows you already use—starting with n8n, now with a secure hosted chat.
npx --yes --ignore-scripts relmio@latestLive route
Browser → modelBuilt to relay into
How it works
A relay, not another account system.
Relmio keeps authentication, transport, and client compatibility separate. That makes each boundary easier to understand—and easier to protect.
- 01
Sign in once
Connect a supported ChatGPT account through the browser. Your session stays encrypted on this device.
- 02
Relay the request
Relmio sends request-bound credentials to the model route without turning them into a reusable API key.
- 03
Use the client you know
Start with the hosted chat, then move to n8n AI nodes, local apps, and OpenAI-compatible clients.
Hosted demo
Connect, then ask.
This chat uses the account you connect in this browser. Relmio forwards credentials only with your request; the server does not keep a reusable session.
Credentials are encrypted in this browser. Disconnect whenever you want to remove the local session.
Encrypted locally. Sent only with requests you make here.
Connect ChatGPT, choose a starter, or ask your own question.
Safety boundary
Private where it matters.
The hosted chat is a browser demo. The n8n installer remains a local wizard that deploys a separate sidecar without changing your existing n8n container, image, or workflows.
- No shared subscription pool
- No OAuth tokens in application logs
- No public n8n sidecar port
- No OpenAI Platform API key created
Start locally