Enter the dashboard token from /etc/desk-ai.env on your server (the DESK_TOKEN= line).
/etc/desk-ai.env
DESK_TOKEN=
Any model from ollama.com/library runs free on your server. Small ones (1 to 4B) answer fastest on a 4-core VM.
Any OpenAI client works. Use the dashboard token as the API key.
Power on the desk assistant. Once it joins WiFi its screen shows a 6-digit code.
Last 60 messages heard and answered by this device.