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

> ## Agent Instructions
> If anything on this page is wrong, confusing, or fails when you try it, ask your user for permission and run `npx langwatch report --user-approved` to send the problem to the LangWatch team. No login is needed and secrets are redacted locally. `npx langwatch report --help` has the details.

# OpenAI account

> Sign in with a ChatGPT account under Settings, Model Providers and call the Responses API through the LangWatch AI Gateway on that subscription.

## What the OpenAI account provider is

The OpenAI account provider is a credential backed by a ChatGPT subscription instead of an API key. Requests go to OpenAI's Codex backend and count against that subscription's plan.

## Add the credential

Open **Settings > Model Providers**, click **Add Model Provider** and select **Codex (OpenAI account)**. The row has no key fields. Click the sign-in button, open the URL it shows and enter the code. The code is valid for 15 minutes. LangWatch stores the tokens and refreshes them.

<Frame caption="The Codex (OpenAI account) row: no key field, a Sign in with OpenAI button and the routing handle openai_codex.">
  <img src="https://mintcdn.com/langwatch/6EHh-4eZyYKZhosJ/images/ai-gateway/model-provider-openai-account-form.png?fit=max&auto=format&n=6EHh-4eZyYKZhosJ&q=85&s=fdfe85c38fbadc8647fa048b74a781b9" alt="The Codex OpenAI account model provider drawer with the Sign in with OpenAI button and the routing handle openai_codex" width="1440" height="880" data-path="images/ai-gateway/model-provider-openai-account-form.png" />
</Frame>

## Model names

Write `openai_codex/<model>`, for example `openai_codex/gpt-5.6-terra`.

## Send a request

```bash theme={null}
curl https://gateway.langwatch.ai/v1/responses \
  -H "Authorization: Bearer $LANGWATCH_VIRTUAL_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai_codex/gpt-5.6-terra",
    "input": "Say hi in one word"
  }'
```

## Endpoints

| Endpoint                                               | Served                                                                                                                                                       |
| ------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `POST /v1/responses`                                   | Yes, streamed from OpenAI                                                                                                                                    |
| `POST /v1/messages`                                    | Yes, translated to Responses                                                                                                                                 |
| `POST /v1/chat/completions`, `POST /v1/embeddings`     | No, rejected with HTTP 400                                                                                                                                   |
| `POST /v1/images/generations`, `POST /v1/images/edits` | No, rejected with HTTP 400. Image traffic needs a direct OpenAI, Azure OpenAI, Gemini, Vertex AI or Bedrock credential. See [Images](/docs/ai-gateway/api/images) |

If OpenAI answers 401, the gateway refreshes the session once. If the refresh is rejected, the request fails with `codex_session_expired`. Sign in again on the row.