> ## 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.

# Assistants

> Use Responses and stored response records instead of an Assistants API.

The hosted API does not clone the OpenAI Assistants API. There are no `/v1/assistants` routes.

Use `/v1/responses` when you need a stateful conversation primitive. The API stores response records for 24 hours and exposes retrieval routes:

| Method | Path                             |
| ------ | -------------------------------- |
| `POST` | `/v1/responses`                  |
| `GET`  | `/v1/responses/{id}`             |
| `GET`  | `/v1/responses/{id}/input_items` |
| `POST` | `/v1/responses/{id}/cancel`      |

## Continue a response

```json theme={null}
{
  "model": "gpt-5.5",
  "previous_response_id": "resp_...",
  "input": [
    { "role": "user", "content": "Continue from the previous answer." }
  ]
}
```

Before the provider call, the gateway hydrates prior messages from the stored response chain.

## Related

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