> ## Documentation Index
> Fetch the complete documentation index at: https://docs.compose.market/llms.txt
> Use this file to discover all available pages before exploring further.

# Text

> Chat completions, Responses-style calls, streaming, tools, and reasoning parameters.

Text models run through chat completions or Responses-style routes.

| Method | Path                            | Notes                                                                      |
| ------ | ------------------------------- | -------------------------------------------------------------------------- |
| `POST` | `/v1/chat/completions`          | OpenAI Chat Completions shape.                                             |
| `POST` | `/v1/responses`                 | Responses-style input, previous response IDs, and stored response records. |
| `POST` | `/external/v1/chat/completions` | OpenAI-compatible external route.                                          |
| `POST` | `/external/v1/responses`        | Responses-style external route.                                            |

## Chat request

```json theme={null}
{
  "model": "gpt-5.5",
  "messages": [
    { "role": "system", "content": "Be concise." },
    { "role": "user", "content": "Write a commit summary." }
  ],
  "stream": true,
  "tools": [],
  "tool_choice": "auto"
}
```

The normalizer preserves common OpenAI and AI SDK fields, including `stream_options`, `parallel_tool_calls`, `response_format`, and reasoning fields where the selected adapter supports them.

## Responses request

```json theme={null}
{
  "model": "gpt-5.5",
  "input": [
    { "role": "user", "content": "Summarize this log." }
  ],
  "stream": false
}
```

Responses can be retrieved by ID after creation.

## Related

* [Tools](/inference/tools)
* [Streaming](/inference/streaming)
* [Inference API reference](/inference/reference/specification)