OpenAI-compatiblestreamingBearer authmodel routing

Reference: Chat Completions

POST /v1/chat/completions — OpenAI-compatible model gateway access through OneAI.

Request

Use provider:model ids from /v1/models. GPT-5 class models use max_completion_tokens.

curl -s https://oneai-saas-api-production.up.railway.app/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_KEY" \
  -d '{
    "model": "openai:gpt-5.2",
    "messages": [
      { "role": "user", "content": "Explain OneAI SaaS in one sentence." }
    ],
    "max_completion_tokens": 300
  }'

Streaming

Set stream: true to receive server-sent events compatible with chat completion chunks.

curl -N https://oneai-saas-api-production.up.railway.app/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_KEY" \
  -d '{
    "model": "openai:gpt-5.2",
    "messages": [
      { "role": "user", "content": "Write a short OneAI launch line." }
    ],
    "stream": true,
    "max_completion_tokens": 300
  }'

Response

OneAI returns standard choices plus oneai.trace for routing observability.

{
  "id": "chatcmpl_...",
  "object": "chat.completion",
  "model": "gpt-5.2-...",
  "provider": "openai",
  "choices": [{ "message": { "role": "assistant", "content": "..." } }],
  "usage": {
    "prompt_tokens": 15,
    "completion_tokens": 57,
    "total_tokens": 72,
    "estimated_cost_usd": 0.00012
  },
  "oneai": {
    "requestId": "chatcmpl_...",
    "trace": { "selectedProvider": "openai", "fallbackUsed": false }
  }
}

Generate

Models