Documentation
Janitor AI
Use NavyAI as the proxy backend for Janitor AI. Point Janitor's custom proxy configuration at the NavyAI chat completions endpoint and use any model from the catalog.
Setup
- Open a chat in Janitor AI and open API Settings
- Select Proxy as the API type
- Choose the Custom proxy configuration
- Set Model name to any NavyAI model, e.g.
claude-sonnet-4.6 - Set the API / proxy URL to
https://api.navy/v1/chat/completions - Paste your NavyAI key (
sk-navy-...) as the API key - Save and send a message to verify the connection
Notes
- Janitor AI expects the full completions URL (
https://api.navy/v1/chat/completions), not just the base URL - The model name must match an ID from
GET /v1/modelsexactly - If replies fail with
403, the model you picked needs a higher plan tier — pick a non-premium model or upgrade - Streaming is supported; leave it enabled for faster perceived responses