Skip to main content

​
OpenAI Chat Completions (HTTP)

Fased’s Gateway can serve a small OpenAI-compatible Chat Completions endpoint. This endpoint is disabled by default. When disabled, the route is not served.
  • POST /v1/chat/completions
  • Same port as the Gateway: http://<gateway-host>:<port>/v1/chat/completions
Under the hood, requests execute as a normal Gateway agent run. Routing, permissions, and config match your Gateway.

​
Authentication

Uses the Gateway auth configuration. Send a bearer token:
  • Authorization: Bearer <token>
Notes:
  • When gateway.auth.mode="token", use gateway.auth.token or FASED_GATEWAY_TOKEN.
  • When gateway.auth.mode="password", use gateway.auth.password or FASED_GATEWAY_PASSWORD.
  • If gateway.auth.rateLimit is configured and too many auth failures occur, the endpoint returns 429 with Retry-After.

​
Choosing an agent

No custom headers required: encode the agent id in the OpenAI model field:
  • model: "fased:<agentId>" (example: "fased:main", "fased:beta")
  • model: "fased/<agentId>" (alias)
  • model: "agent:<agentId>" (alias)
Or target a specific Fased agent by header:
  • x-fased-agent-id: <agentId> (default: main)
  • x-fased-agent: <agentId> (alias)
Advanced:
  • x-fased-session-key: <sessionKey> to fully control session routing.

​
Enabling the endpoint

Set gateway.http.endpoints.chatCompletions.enabled to true:

​
Disabling the endpoint

Set gateway.http.endpoints.chatCompletions.enabled to false:

​
Session behavior

By default, the endpoint is stateless per request. A new session key is generated for each call. If the request includes an OpenAI user string, the Gateway derives a stable session key from it so repeated calls can share an agent session.

​
Message support

  • system and developer messages become extra system instructions for the run.
  • user, assistant, tool, and function messages become conversation context.
  • Text content is supported. Non-text message parts are ignored by this compatibility endpoint.
  • Request bodies are capped at 1 MiB.
  • Usage fields are returned in the OpenAI shape, but token counts are currently 0.

​
Streaming (SSE)

Set stream: true to receive Server-Sent Events (SSE):
  • Content-Type: text/event-stream
  • Each event line is data: <json>
  • Stream ends with data: [DONE]

​
Examples

Non-streaming:
Streaming: