LLM providers
API keys live on the inference server at
/etc/sentinel/llm-keys.env. Saving here updates that file and restarts the inference server (~5s). Keys are never displayed back.loading…
rolling 24h spend
loading…
rolling 24h window · auto-refreshes every 30s
OpenAI
not configured
Anthropic
not configured
Gemini
not configured
After save: the inference server restarts and refreshes the live model list for that provider within a few seconds. If you set a wrong key, the provider will show
configured: true but calls will fail at runtime — check journalctl -u inference-server -n 20 on the box if calls fail right after.