Create a chat completion

POST

/v1/chat/completions

Create a chat completion. Model IDs infer the provider; no provider extension is needed for standard clients. Claude and Gemini support tool calls and streaming. For streams, failures before the first event return HTTP 502; failures after output begins emit a data event with an error envelope and no [DONE]. Treat incomplete streams as failures. Request IDs are returned in headers. Usage and cost attribution use server-controlled X-User-ID and optional session headers.

Request

Header Params

Body Params application/jsonRequired

Examples

Responses

🟢200

application/json

Non-streaming — Content-Type: application/json

Streaming — Content-Type: text/event-stream. Each event:
data: {StreamChunk json}\n\n. Ends with data: [DONE]\n\n.

Headers

Body

🟠400BadRequest

🟠401Unauthorized

🟠429RateLimited

🔴502ProviderError

🔴503Unavailable

Request Example

Shell

JavaScript

Java

Swift

cURL

Plain text questionFunction callingStreaming response

curl --location '/v1/chat/completions' \
--header 'X-User-ID;' \
--header 'X-Session-ID;' \
--header 'Content-Type: application/json' \
--data '{
    "provider": "claude",
    "model": "claude-sonnet-4-6",
    "messages": [
        {
            "role": "user",
            "content": "Explain quantum entanglement in one paragraph."
        }
    ],
    "max_tokens": 256
}'

Response Example

200 - Example 1

{
    "id": "chatcmpl-1234567890",
    "object": "chat.completion",
    "created": 0,
    "model": "claude-sonnet-4-6",
    "provider": "claude",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "system",
                "content": "string",
                "name": "string",
                "tool_calls": [
                    {
                        "id": "call_abc123",
                        "type": "function",
                        "function": {
                            "name": "get_weather",
                            "arguments": "{\"location\":\"Paris, FR\"}"
                        },
                        "index": 0
                    }
                ],
                "tool_call_id": "call_abc123"
            },
            "finish_reason": "stop",
            "logprobs": null
        }
    ],
    "usage": {
        "prompt_tokens": 12,
        "completion_tokens": 48,
        "total_tokens": 60
    }
}

Modified at 2026-10-20 19:28:20

Operations and observability

Create text embeddings