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

# Quick Start

> Connect your AI agent to Borough in 2 minutes

## Prerequisites

1. A Borough account
2. Optional: a Borough API key if you want paid-tier tools
3. An MCP-compatible AI client

If you are connecting from Claude or Claude Desktop, use the dedicated guide:

* [Add Borough to Claude](/borough/mcp/add-to-claude)

## Remote Connection (Recommended)

Most MCP clients support connecting to remote servers directly. Use the Borough MCP endpoint:

```
https://borough.qwady.app/mcp
```

For clients without OAuth support, pass your Borough API key via the `Authorization` header:

```
Authorization: Bearer BOROUGH-<your_key>
```

## Client Setup

<Tabs>
  <Tab title="Claude Connectors">
    Borough supports an OAuth-based connector flow for Claude custom connectors.

    1. In Claude, go to **Settings → Connectors**
    2. Click **Add custom connector**
    3. Enter the URL: `https://borough.qwady.app/mcp`
    4. Claude auto-discovers the OAuth flow and opens the consent page
    5. Enter your Borough API key to enable paid tools, or continue without a key for the free Borough MCP tool set
    6. Click **Authorize**

    Claude can now query NYC real estate data directly.

    <Note>See [Add Borough to Claude](/borough/mcp/add-to-claude) for the exact current flow, including free-plan and team/org setup.</Note>
  </Tab>

  <Tab title="Claude Desktop">
    Claude Desktop supports the same remote custom connector flow as Claude web.

    1. Open **Settings → Connectors**
    2. Click **Add custom connector**
    3. Enter `https://borough.qwady.app/mcp`
    4. Complete the Borough OAuth flow

    If you need the full step-by-step path, use:

    * [Add Borough to Claude](/borough/mcp/add-to-claude)
  </Tab>

  <Tab title="Cursor">
    Add to `.cursor/mcp.json` in your project:

    ```json theme={null}
    {
      "mcpServers": {
        "borough": {
          "url": "https://borough.qwady.app/mcp",
          "headers": {
            "Authorization": "Bearer BOROUGH-<your_key>"
          }
        }
      }
    }
    ```
  </Tab>

  <Tab title="VS Code">
    Add to `.vscode/mcp.json` in your project:

    ```json theme={null}
    {
      "servers": {
        "borough": {
          "type": "http",
          "url": "https://borough.qwady.app/mcp",
          "headers": {
            "Authorization": "Bearer BOROUGH-<your_key>"
          }
        }
      }
    }
    ```
  </Tab>

  <Tab title="npx">
    Run locally via the [`@borough/mcp`](https://www.npmjs.com/package/@borough/mcp) stdio bridge:

    ```bash theme={null}
    BOROUGH_API_KEY=BOROUGH-<your_key> npx @borough/mcp
    ```

    This spawns a local stdio-to-HTTP bridge that connects to the remote MCP server.
  </Tab>
</Tabs>

## ChatGPT (GPT Actions)

ChatGPT does not support MCP natively. Instead, you can create a custom GPT with an Action that calls the Borough REST API directly using the [Borough OpenAPI spec](https://qwady.wiki/borough/openapi.yaml):

1. Go to [ChatGPT GPT Builder](https://chatgpt.com/gpts/editor)
2. Under **Actions**, click **Create new action**
3. Import the OpenAPI spec from the URL above (or paste the YAML)
4. Under **Authentication**, select **API Key** with header `Authorization` and prefix `Bearer`
5. Enter your Borough API key as the value

## Test It

Once connected, try asking your AI agent:

* "Search for no-fee 1-bedroom apartments in Manhattan under \$3,500"
* "List all Brooklyn neighborhoods"
* "What's the median rent in the East Village?"

The agent will automatically select the right Borough tools and return formatted results.

## Troubleshooting

If tools aren't showing up:

1. Verify your Borough account or API key is valid by visiting the [customer portal](https://polar.sh/qwady-solutions-llc/portal)
2. Check that the MCP endpoint is reachable: `curl https://borough.qwady.app/mcp` (a `401` response is expected and confirms it's up; the endpoint requires authentication)
3. Restart your AI client after configuration changes
4. See the [troubleshooting guide](/borough/mcp/troubleshooting) for more help