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

# Connect Codex

> Configure the traxy MCP connection in Codex and complete OAuth authorization.

Use a Streamable HTTP MCP connection to work with traxy from Codex. The connection uses browser OAuth, not a traxy REST API key.

## Connect from the CLI

With Codex installed, add the server:

```bash theme={null}
codex mcp add traxy --url https://api.traxy.ai/api/v1/mcp
```

Complete browser authorization if it opens. If authentication is still required, run:

```bash theme={null}
codex mcp login traxy
```

Choose the default workspace, any additional workspaces, and the actions you want to grant. Return to Codex when authorization completes.

## Connect from a desktop or IDE client

Open the client's MCP server settings, add a **Streamable HTTP** server, and use the same URL. Save it and use **Authenticate** when offered. Leave bearer-token and custom authorization-header settings empty for this OAuth connection.

Codex clients on the same host can share MCP configuration. ChatGPT web uses its own plugin connection flow; a local CLI configuration does not connect the web client. See [OpenAI's MCP setup guide](https://learn.chatgpt.com/docs/extend/mcp) for the current client-specific steps.

## Verify and use the connection

Run `codex mcp list` to inspect configured servers. In a conversation, ask Codex to list the authorized traxy workspaces and read a few leads from the one you select.

Review [workspace selection](https://docs.traxy.ai/api-reference/mcp/workspaces.md) before using several workspaces and [permissions](https://docs.traxy.ai/api-reference/authentication.md) before requesting writes.


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