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

# Connect Codex to Corunner MCP

> Configure Corunner Engineering Memory MCP in Codex CLI and the Codex IDE extension.

Codex shares MCP configuration between the CLI and the IDE extension. Add Corunner once, then complete the browser-based OAuth flow when Codex connects.

## Add Corunner

Run:

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

Codex prompts you to authenticate when it first connects. Sign in, choose the Corunner workspace, and approve the scopes your work requires.

## Configure manually

You can configure the server in `~/.codex/config.toml`:

```toml theme={null}
[mcp_servers.corunner]
url = "https://api.corunner.ai/mcp"
auth = "oauth"
oauth_resource = "https://api.corunner.ai/mcp"
scopes = [
  "engineering:context:read",
  "engineering:decisions:read",
  "engineering:requirements:read",
  "engineering:history:read",
  "engineering:incidents:read",
  "engineering:change:validate",
]
```

Request only the scopes needed for your workflow. For example, use `engineering:context:read` and `engineering:change:validate` for context retrieval and final change checks.

## Verify the connection

Run `codex mcp list` and confirm that `corunner` is listed. Then ask Codex to retrieve context for a repository file. Include the repository's canonical Git remote so Corunner can resolve the correct workspace repository.

<Warning>
  Do not commit `~/.codex/config.toml` or copy access tokens into a project configuration file.
</Warning>

For the latest Codex command syntax, see the [OpenAI MCP setup guide](https://developers.openai.com/learn/docs-mcp).


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