> For the complete documentation index, see [llms.txt](https://help.cubelet.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.cubelet.ai/platform-and-integration/connecting-a-client/cursor.md).

# Cursor

*Connect Cursor to a cubelet.ai MCP server via `.cursor/mcp.json`.*

## Configure the server

Add an entry under `mcpServers` in either:

* `.cursor/mcp.json` at your repo root — scoped to that project.
* `~/.cursor/mcp.json` in your home directory — available in every project.

A remote MCP server takes a `url` and an optional `headers` object; no `command` is needed. The cubelet.ai fleet authenticates with a bearer token (see Auth & access):

{% code collapsedlinecount="10" %}

```json
{
  "mcpServers": {
    "cubelet-manufacture": {
      "url": "https://manufacture.cubelet.ai/mcp",
      "headers": {
        "Authorization": "Bearer ${env:CUBELET_MCP_TOKEN}"
      }
    }
  }
}
```

{% endcode %}

Replace the URL with the `/mcp` URL of the server you picked from the MCP servers page.

Cursor resolves `${``env:NAME}` references in `url` and `headers` from your shell environment when it reads the config. Keep the actual credential value out of the file itself — especially in the project-scoped `.cursor/mcp.json`, which is typically checked into the repo.