# Connect your AI app to AgenticRelay with MCP

> Add the AgenticRelay MCP server to Claude, ChatGPT, Cursor or VS Code. One sign-in reaches every workspace you belong to. Not billed, with rate limits.

Source: https://agenticrelay.app/mcp

Set up the MCP server in Claude, ChatGPT, Cursor or VS Code with one URL.

## The server URL

The AgenticRelay MCP server is available to every workspace. Give your AI app this URL and nothing else. It finds the sign-in and token endpoints on its own.

```
https://mcp.agenticrelay.app/mcp
```

## One connection, every workspace

A connection belongs to your account, not to a workspace. You connect once, and the app reaches every workspace where you are an active member, with the role you have in each one. A workspace you join later works without reconnecting.

Every tool takes an optional workspaceId. When your account has several workspaces, the AI calls list_workspaces to see their ids and your role in each, then passes the right workspaceId. With a single workspace it can leave it out.

Access is checked on every call. If you leave a workspace or lose a role, the app loses it on its next call.

## Add it to your AI app

Use the same server URL everywhere. Pick your app:

### Claude (claude.ai and Claude Desktop)

1. Open Settings, then Connectors.
2. Choose Add custom connector.
3. Paste the server URL and confirm.
4. Sign in to AgenticRelay and allow access when asked.

### Claude Code

Run this command in your terminal, then use /mcp inside Claude Code to sign in:

```
claude mcp add --transport http agenticrelay https://mcp.agenticrelay.app/mcp
```

### ChatGPT

1. Open Settings, then Apps, then Advanced settings.
2. Turn on Developer mode.
3. Choose Create app, paste the server URL and choose OAuth.
4. Sign in to AgenticRelay and allow access when asked.

### Cursor

Cursor cannot register with the server directly yet, so it connects through the mcp-remote bridge. Add this to your mcp.json:

```
{
  "mcpServers": {
    "agenticrelay": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.agenticrelay.app/mcp"
      ]
    }
  }
}
```

### VS Code

Open the Command Palette, run MCP: Add Server, choose HTTP and paste the server URL. Or add this to your mcp.json:

```
{
  "servers": {
    "agenticrelay": {
      "type": "http",
      "url": "https://mcp.agenticrelay.app/mcp"
    }
  }
}
```

## What happens when you connect

Your AI app registers itself, then sends you to a AgenticRelay consent page. Sign in if needed and review what it asks for. The page lists the workspaces the app will be able to work in, with your role in each. You approve once.

The server speaks OAuth 2.1 with PKCE and dynamic client registration, so no API key is shared and nothing is copied by hand. The app only does what your role allows in the workspace a call targets.

## Revoke access anytime

In AgenticRelay, open Settings and then Connected apps, in the Account group. Disconnect an app and its access ends in every workspace on its next call.

## Cost and limits

Using the MCP server is not billed. We charge for what your AI employees use, as on any plan.

Each connection has anti-abuse rate limits of about 120 calls and 20 writes per 60 seconds. MCP calls count against the rate limit, and a call over it gets a 429 with a Retry-After header.