Skip to content

api.md

Latest commit

 

History

History
148 lines (120 loc) · 3.65 KB

api.md

File metadata and controls

148 lines (120 loc) · 3.65 KB

API Usage

← Back to README

Supported API shapes

  • POST /v1/messages — Anthropic Messages API
  • POST /v1/chat/completions — OpenAI Chat Completions API
  • POST /v1/responses — OpenAI Responses API
  • GET /v1/models — OpenAI models API
  • GET /models — compatibility alias for /v1/models

All of these routes reuse the same multi-account pool, file upload pipeline, tool bridge, and failover logic.

/v1/responses is currently stateless, so previous_response_id is not supported.

Anthropic Messages request

/v1/messages accepts Anthropic Messages API payloads.

curl http://localhost:3000/v1/messages \
  -H "x-api-key: <api_key>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "opus-4.6",
    "max_tokens": 1024,
    "messages": [
      { "role": "user", "content": "Describe the main components of this project." }
    ]
  }'

If model is omitted, the service falls back to proxy.default_model.

OpenAI Models request

curl http://localhost:3000/v1/models \
  -H "Authorization: Bearer <api_key>"

The response returns normalized friendly model IDs that can be passed back into /v1/chat/completions, /v1/responses, or /v1/messages.

GET /models returns the same payload for clients that probe the bare root alias.

OpenAI Chat Completions request

curl http://localhost:3000/v1/chat/completions \
  -H "Authorization: Bearer <api_key>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "messages": [
      { "role": "user", "content": "Summarize the architecture of notion-manager." }
    ]
  }'

Supported request features include messages, stream, tools, tool_choice, response_format, and inline file/image inputs.

OpenAI Responses request

curl http://localhost:3000/v1/responses \
  -H "Authorization: Bearer <api_key>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "input": "List the main subsystems in this project."
  }'

Supported request features include input, instructions, stream, tools, tool_choice, text.format, and inline file/image inputs.

Search overrides

curl http://localhost:3000/v1/messages \
  -H "x-api-key: <api_key>" \
  -H "Content-Type: application/json" \
  -H "X-Web-Search: true" \
  -H "X-Workspace-Search: false" \
  -d '{
    "model": "sonnet-4.6",
    "messages": [
      { "role": "user", "content": "Search for recent information about Go 1.25." }
    ]
  }'

File uploads

Supported media types:

  • image/png
  • image/jpeg
  • image/gif
  • image/webp
  • application/pdf
  • text/csv
curl http://localhost:3000/v1/messages \
  -H "x-api-key: <api_key>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "sonnet-4.6",
    "max_tokens": 600,
    "messages": [{
      "role": "user",
      "content": [
        {
          "type": "document",
          "source": {
            "type": "base64",
            "media_type": "application/pdf",
            "data": "<base64>"
          }
        },
        {
          "type": "text",
          "text": "Summarize this PDF."
        }
      ]
    }]
  }'

Research mode

curl -N http://localhost:3000/v1/messages \
  -H "x-api-key: <api_key>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "researcher",
    "stream": true,
    "max_tokens": 16000,
    "thinking": { "type": "enabled", "budget_tokens": 50000 },
    "messages": [
      { "role": "user", "content": "Map common architectural patterns used by Notion AI proxy tools." }
    ]
  }'

Research mode is single-turn, ignores file uploads, ignores custom tools, and runs with a longer timeout path.