DISTLM · LAN
Register this browser on signaling. Set your public model URL (ngrok, LAN IP, etc.) so the host can call your OpenAI-compatible /v1/chat/completions. When the host runs a job, you will see the prompt here.
If several logical agents share this mesh connection, give each a distinct id so the host feed can show them separately (peer_id#instance_id). Verify counts with GET /jobs/<job_id>/response-summary on the orchestrator.
Saved in this browser (localStorage). After connecting you can edit and click "Push URL to server".
Session override (sessionStorage).