> For the complete documentation index, see [llms.txt](https://docs.emseapea.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.emseapea.ai/builder-guide/connect-your-ai.md).

# Connect emseapea to your AI tool

emseapea speaks MCP — the open protocol Claude, ChatGPT, Cursor, and Copilot all understand. Connecting takes one URL and one sign-in.

1. Add your organization's emseapea MCP server URL to your AI tool (e.g. in Claude: add a custom connector with `https://<your-org-host>/mcp`).
2. Your tool opens a sign-in window — use your normal emseapea account. The consent screen names the organization you're connecting for.
3. Done. Your AI now has four governed tools:

| Tool                   | What it does                                                          |
| ---------------------- | --------------------------------------------------------------------- |
| `request_app`          | Start a governed build: creates the intent record, pending approval   |
| `check_request_status` | Where your request is in the queue                                    |
| `get_workspace`        | After approval: your repo URL, credential handle, approved connectors |
| `list_my_apps`         | Your apps and their states                                            |

## Two connections, two jobs

Your AI ends up with **two** MCP connections that do different jobs:

* **emseapea** — governance: requesting, tracking, and receiving workspaces.
* **your company's gateway** — data: the connection that actually reaches ERP, CRM, or HR, unlocked by the credential emseapea issues you on approval.

emseapea never proxies your company's data; it arranges your access to it.

No MCP-capable tool? The web form does exactly the same thing — see [Requesting an app](/builder-guide/requesting-an-app.md).
