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

# Responses API

> Send a text request using the OpenAI Responses format.

APIONE provides `POST /v1/responses`. Choose a model that supports Responses. A Chat Completions request body cannot be reused unchanged.

## Minimal request

Set `APIONE_API_KEY` and `APIONE_MODEL` as described in the [quickstart](https://apione.ai/docs/en/quickstart.md). Choose an ID from [available models](https://apione.ai/docs/en/models/catalog.md). This example uses the basic Responses text format:

```bash theme={null}
curl --fail-with-body --max-time 120 https://api.apione.ai/v1/responses \
  -H "Authorization: Bearer ${APIONE_API_KEY:?Set APIONE_API_KEY first}" \
  -H "Content-Type: application/json" \
  --data "{\"model\":\"${APIONE_MODEL:?Set APIONE_MODEL first}\",\"input\":\"Hello\"}"
```

Read `output[]`: select items with `type: "message"`, then text blocks with `type: "output_text"` in their `content[]`. The text is in `text`, rather than Chat Completions' `choices[0].message.content`.

## Configure the client

Use `https://api.apione.ai/v1` as the OpenAI SDK Base URL. For clients requiring a complete endpoint, use `https://api.apione.ai/v1/responses`. See [Codex integration](https://apione.ai/docs/en/guides/codex.md).

Verify text before tools or streaming. Storage, background tasks, built-in tools, and extension fields require separate compatibility checks. On failure, save the status, error body, and the actual request-identifier header names and values for [troubleshooting](https://apione.ai/docs/en/support/errors.md).

Sources: [APIONE endpoints](https://apione.ai/docs/en/api/endpoints.md) and the [OpenAI Responses format](https://developers.openai.com/api/reference/resources/responses/methods/create).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.