API reference

Chat Completions

Send messages and receive text, tool calls, or structured data.

On this page

Request
Messages
Response
Model-specific compatibility

Request

POST/chat/completions

Creates a response using a model available to the key.

cURL

curl https://code.verboo.ai/router/v1/chat/completions \
  -H "Authorization: Bearer $VERBOO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<MODEL_ID>",
    "messages": [{"role": "user", "content": "Explain why low latency matters for developer tools."}]
  }'

Field	Required	Description
model	yes	An ID returned by GET /models.
messages	yes	Ordered message history.
stream	no	When true, returns SSE events.
temperature	no	Sampling; support and range depend on the provider.
max_tokens / max_completion_tokens	no	Output limit, depending on the model.
tools / tool_choice	no	Functions available to the model.
reasoning_effort	no	Only for models with reasoning.
response_format	no	JSON object or JSON schema; may depend on the plan.

Messages

Role	Use
system	Defines global behavior and constraints.
user	User input; may contain text or multimodal parts.
assistant	A previous answer or tool call request.
tool	Function result associated through tool_call_id.

Response

200 · application/json

{
  "id": "chatcmpl_...",
  "object": "chat.completion",
  "created": 1786752000,
  "model": "model-id",
  "choices": [{
    "index": 0,
    "message": {"role": "assistant", "content": "Low latency keeps feedback loops short."},
    "finish_reason": "stop"
  }],
  "usage": {"prompt_tokens": 18, "completion_tokens": 10, "total_tokens": 28}
}

Read choices[0].message.content for text, choices[0].message.tool_calls for functions, and usage for measured consumption.

Model-specific compatibility

Not every field is universal

The router normalizes the core flow, but providers may ignore or reject specific parameters. Use the guides and test your selected model before production.