Before you start

1. Go to Settings > Integrations > MCP
2. Click Create. A confirmation modal will display the new key
3. Copy the Endpoint (URL) from the table — you'll paste this into your AI tool
   

1. Click the settings gear next to your key and choose Tools to enable or disable individual capabilities for this key
   

Caution: Treat your MCP key like a password. It can trigger tools connected to your account and access sensitive information.

Claude

Setting up Edbound MCP in Claude

1. Navigate to Claude > Settings
   
2. Click on Connectors > Add custom connector
3. Enter a name and paste the Endpoint (URL) from your Edbound account
   

1. Click Add
   

Accessing Edbound MCP tools in Claude

 

1. Open a new chat in Claude
2. Click the + (plus) button next to the message input box
3. Select Connectors from the menu
4. In the connectors panel, enable the Edbound toggle
5. Start your prompt and ask Claude to perform actions using the enabled Edbound tools

Note: You only need to enable Edbound MCP once. After setup, Edbound tools will be available in all your Claude chats.

Cursor

Setting up Edbound MCP in Cursor

 

1. Open Cursor settings (⇧ + ⌘ + J on Mac or Ctrl + Shift + J on Windows)
2. Go to the MCP Tools tab and click New MCP Server
3. Copy and paste the following JSON configuration, replacing YOUR_ENDPOINT_URL with the Endpoint (URL) from your Edbound account:
   

{
  "mcpServers": {
    "EdboundMCP": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "YOUR_ENDPOINT_URL",
        "--transport",
        "http-only"
      ]
    }
  }
}

1. Save the configuration using ⌘ + S or Ctrl + S

Note: Make sure the chat is set to Agent mode to access Edbound MCP tools within Cursor.

Windsurf

Setting up Edbound MCP in Windsurf

1. Open Windsurf Settings (⌘ + , on Mac or Ctrl + , on Windows)
2. In the left sidebar, click Cascade
3. Click Open MCP Registry
4. In the MCP Registry screen, click the ⚙ Settings (gear) icon
   

1. This opens the mcp_config.json file
   
2. Paste the MCP server JSON configuration into this file, replacing YOUR_ENDPOINT_URL with your actual Endpoint (URL):
   
   

{
  "mcpServers": {
    "EdboundMCP": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "YOUR_ENDPOINT_URL",
        "--transport",
        "http-only"
      ]
    }
  }
}

1. Save the file
2. After saving, the MCP server will be available in Cascade, and you can start using it in your agent/chat

VS Code

Setting up Edbound MCP in Visual Studio Code

1. Open the Visual Studio Code Command Palette (⇧ + ⌘ + P on Mac, Ctrl + Shift + P on Windows)
2. Type "MCP: Add Server…" and press Enter
3. Choose "HTTP (HTTP or Server-Sent Events)" and press Enter
4. Paste the Endpoint (URL) from your Edbound account into the "Server URL" field and press Enter
5. Provide a server name (e.g., "Edbound") and press Enter
6. In the chat, type your prompt and ask GitHub Copilot to perform actions using the enabled Edbound tools

Note: You must have GitHub Copilot enabled and set to Agent mode in Visual Studio Code for Edbound MCP to function correctly.

Other tools

1. Endpoint URL: Copy the Endpoint (URL) from your Edbound account
2. Transport Type: HTTP (Server-Sent Events)

Generic setup steps

1. Open your AI tool's settings or configuration panel
   
1. Look for MCP, Integrations, or External Tools section
2. Add a new MCP server connection
3. Enter a name (e.g., "Edbound")
4. Paste the Endpoint (URL)
5. Select HTTP or SSE (Server-Sent Events) as the transport type if prompted
6. Save the configuration
7. Enable or activate the MCP connection in your AI tool
   

1. For tools using JSON configuration
   
1. If your tool requires a JSON configuration file, use this template:
   

{
  "mcpServers": {
    "EdboundMCP": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "YOUR_ENDPOINT_URL",
        "--transport",
        "http-only"
      ]
    }
  }
}

1. Replace YOUR_ENDPOINT_URL with the Endpoint (URL) from your Edbound account.

Troubleshooting

Frequently Asked Questions