API 参考

Fathom REST API — 完整参考。

Base URL

https://api.wyrmlabs.com/v1

Authentication

All API requests require authentication via an API key passed in the Authorization header:

Authorization: Bearer <your_api_key>

API keys are issued through the Wyrmlabs dashboard. Enterprise customers can configure custom authentication schemes including mutual TLS and OAuth 2.0.

Chat Completions

Generate a model response for a chat conversation.

POST /chat/completions

Request Body

{
  "model": "fathom-1.0",
  "messages": [
    {"role": "system", "content": "You are Fathom, an AI assistant."},
    {"role": "user", "content": "Hello, how are you?"}
  ],
  "temperature": 0.7,
  "max_tokens": 2048,
  "stream": false
}

Parameters

model (required) — Model identifier: "fathom-1.0" or "fathom-1.0-mini"
messages (required) — Array of message objects with role and content
temperature (optional, default: 0.7) — Sampling temperature (0.0–2.0)
max_tokens (optional, default: 2048) — Maximum tokens in the response
stream (optional, default: false) — Enable server-sent events streaming
top_p (optional, default: 1.0) — Nucleus sampling parameter
stop (optional) — Custom stop sequences (array of strings)

Streaming

Set stream: true in the request body to enable server-sent events (SSE) streaming. Each event is a JSON object prefixed with "data: ". The stream ends with "data: [DONE]". Streaming is recommended for interactive applications to reduce perceived latency.

Rate Limits

Free Tier: 60 requests/min, 10,000 tokens/min

API Tier: 1,000 requests/min, 100,000 tokens/min

Enterprise: Custom rate limits, no hard caps

Rate limit headers are included in all API responses: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset.

Error Codes

400 — Bad Request: Invalid input parameters

401 — Unauthorized: Invalid or missing API key

429 — Rate Limited: Exceeded rate limit

500 — Internal Server Error: Contact support if persistent

503 — Service Unavailable: Model loading or temporary overload

For the complete API reference including embeddings, fine-tuning endpoints, and batch processing, see our full documentation.