A private, on-premise AI assistant that runs on your own hardware — no cloud.
Starting up…
By default this page uses this origin's /api/chat endpoint, which the app server proxies to the model. Set a custom URL to connect directly to an OpenAI-compatible server, or switch to demo mode for canned replies.
/api/chat
Requests go to {url}/v1/chat/completions. Blank uses the same-origin /api/chat endpoint. A trailing /v1 is accepted and normalized. Saved in this browser only.
{url}/v1/chat/completions
/v1
Blank uses the first model the backend reports.
