Skip to content

API.md

Latest commit

 

History

History
146 lines (108 loc) · 2.72 KB

API.md

File metadata and controls

146 lines (108 loc) · 2.72 KB
Raw

API Reference

Base URL: http://<host>:<port> (default http://127.0.0.1:3040)

Authentication header required on all /v1/* endpoints:

Authorization: Bearer <aibridge-api-key>

GET /health

Health probe endpoint (no auth required).

Response

{
  "status": "ok",
  "platform": "ide",
  "copilot": "enabled"
}

GET /v1/models

Returns model list in OpenAI-compatible format.

Response

{
  "object": "list",
  "data": [
    {
      "id": "gpt-4o",
      "object": "model",
      "created": 1710000000,
      "owned_by": "github-copilot"
    }
  ]
}

Notes:

  • Uses discovered Copilot models when available.
  • Returns fallback list when discovery fails or no project is open.

GET /v1/models/{id}

Returns a single model entry or 404.

POST /v1/chat/completions

OpenAI-compatible chat completions endpoint.

Request (typical)

{
  "model": "gpt-4o",
  "messages": [
    {"role": "system", "content": "You are concise."},
    {"role": "user", "content": "Tell me a short joke about Java."}
  ],
  "stream": true,
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "read_file",
        "description": "Read a local file",
        "parameters": {"type": "object"},
        "strict": true
      }
    }
  ]
}

Supported request fields

  • model: string?
  • messages: ChatMessage[] (required)
  • stream: boolean (default false)
  • tools: ChatTool[]?
  • tool_choice: any?
  • max_tokens: int?
  • temperature: double?

Compatibility notes:

  • Unknown request fields are ignored.
  • tools[].function.strict is accepted.
  • messages[].content can be string or array-like OpenAI content blocks; text is flattened.

Streaming response

SSE chunks via text/event-stream:

data: {"id":"chatcmpl-...","object":"chat.completion.chunk","choices":[...]}

data: {"id":"chatcmpl-...","object":"chat.completion.chunk","choices":[{"delta":{},"finish_reason":"stop"}]}

data: [DONE]

A terminal chunk with empty delta and non-null finish_reason is expected OpenAI-style behavior.

Non-streaming response

OpenAI-style chat.completion JSON with one choice.

POST /v1/completions

Legacy endpoint.

Request:

{
  "model": "gpt-4o",
  "prompt": "Write one sentence.",
  "stream": false
}

Behavior:

  • Converts prompt payload into chat format and routes to the same pipeline as /v1/chat/completions.

Error responses

Typical error JSON:

{"error": "Invalid API Key"}

Common status codes:

  • 400: invalid/empty request body
  • 401: invalid bearer key
  • 404: model not found
  • 429: rate limit exceeded
  • 503: no project / server capacity / Copilot unavailable
  • 504: request timeout
  • 500: internal error