Integrations

Network modes, health checks, additional routing and API integration model.

Network modes

01Public endpoint.
02Public endpoint + corporate IP allowlist.
03Site-to-site VPN / private network.

Changing network mode does not require changing application API paths.

Health & readiness

GET /health/live
GET /health/ready
GET /v1/models
GET /v1/balance

/health/ready checks the controlled gateway and key internal dependencies; it does not guarantee a specific external LLM provider.

Additional route model

AI Gate can be an additional route. The primary route and fallback execution remain client-controlled.

API compatibility

Inference uses one API gateway with OpenAI- and Anthropic-compatible formats. The production endpoint depends on the selected network architecture.