Connect RightMessage to AI Assistants via MCP

RightMessage's MCP (Model Context Protocol) server lets AI assistants like Claude, ChatGPT, and Cursor manage your personalization setup directly. Your AI can create campaigns, build flows, analyze performance, and export data — all through natural conversation.

You'll need an API token and a few minutes to configure your AI client.

Generate your API token

First, create a token that your AI assistant will use to access your RightMessage account.

Go to your Profile (click your avatar, then "Your Profile")
Switch to the API Tokens tab
Enter a name like "Claude Desktop" or "Cursor IDE"
Click Create
Copy the token immediately — you won't see it again

Keep your API token secure. Anyone with this token can modify your RightMessage projects.

Configure your AI client

Each AI assistant stores its MCP configuration in a different location. Find your client below and follow the setup steps.

Claude Code (CLI)

Claude Code is Anthropic's command-line tool with built-in MCP support. If you have the CLI installed, adding RightMessage takes one command.

In your terminal, run:

claude mcp add --transport http rightmessage https://api.rightmessage.com/mcp/rightmessage --header "Authorization: Bearer YOUR_API_TOKEN_HERE"

Replace YOUR_API_TOKEN_HERE with the token you copied earlier.

Verify the server was added:

claude mcp list

You should see rightmessage in the list. Now when you use Claude Code, ask "Show me my RightMessage projects" to test the connection.

Use --scope project to add the server only to your current project instead of globally. Add --env VAR=value to pass environment variables.

Claude Desktop

Claude Desktop reads its configuration from a JSON file on your computer.

Open Claude Desktop
Go to Settings → Developer → Edit Config (this opens the config file in your text editor)
Add RightMessage to the mcpServers section:

{
  "mcpServers": {
    "rightmessage": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://api.rightmessage.com/mcp/rightmessage",
        "--header",
        "Authorization: Bearer YOUR_API_TOKEN_HERE"
      ]
    }
  }
}

Replace YOUR_API_TOKEN_HERE with the token you copied earlier
Save the file
Fully quit and restart Claude (right-click the dock icon and select Quit)

You'll see a hammer 🔨 icon in the chat window when RightMessage tools are available. Try asking "Show me my RightMessage projects" to test the connection.

Claude Desktop uses mcp-remote to connect to HTTP-based MCP servers. The first time you use it, npx will download the package automatically.

Cursor

Cursor has native support for HTTP-based MCP servers.

In Cursor, open Settings (Cmd+, on Mac, Ctrl+, on Windows)
Search for "MCP" in settings
Click Edit in settings.json
Add RightMessage to the mcpServers section:

{
  "mcpServers": {
    "rightmessage": {
      "url": "https://api.rightmessage.com/mcp/rightmessage",
      "headers": {
        "Authorization": "Bearer YOUR_API_TOKEN_HERE"
      }
    }
  }
}

Replace YOUR_API_TOKEN_HERE with your token
Save and restart Cursor

Toggle RightMessage tools on/off in the Cursor chat panel. Ask "What campaigns are running in my RightMessage account?" to verify.

Cursor supports environment variables in its config. See the Cursor MCP docs for the exact syntax for referencing your token via an env var instead of hardcoding it.

VS Code with GitHub Copilot

GitHub Copilot supports MCP through VS Code's extension system.

Open VS Code
Press Cmd+Shift+P (Mac) or Ctrl+Shift+P (Windows/Linux)
Type "MCP: Add Server" and select it
Choose HTTP as the transport type
Enter the URL: https://api.rightmessage.com/mcp/rightmessage
Add the authorization header:
- Header name: Authorization
- Header value: Bearer YOUR_API_TOKEN_HERE
Save the configuration

Alternatively, create a .vscode/mcp.json file in your workspace:

{
  "mcpServers": {
    "rightmessage": {
      "url": "https://api.rightmessage.com/mcp/rightmessage",
      "headers": {
        "Authorization": "Bearer YOUR_API_TOKEN_HERE"
      }
    }
  }
}

Switch Copilot to Agent mode in the chat panel to access MCP tools.

ChatGPT Desktop

ChatGPT Desktop uses a similar configuration to Claude.

Locate your ChatGPT configuration file:
- macOS: ~/Library/Application Support/ChatGPT/mcp_config.json
- Windows: %APPDATA%\ChatGPT\mcp_config.json
- Linux: ~/.config/ChatGPT/mcp_config.json
Create or edit the file with this configuration:

{
  "mcpServers": {
    "rightmessage": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://api.rightmessage.com/mcp/rightmessage",
        "--header",
        "Authorization: Bearer YOUR_API_TOKEN_HERE"
      ]
    }
  }
}

Replace YOUR_API_TOKEN_HERE with your token
Save and restart ChatGPT Desktop

Google Gemini (CLI)

Gemini's command-line tools support MCP server configuration.

Open your terminal
Run the Gemini MCP add command:

gemini mcp add --transport http --trust https://api.rightmessage.com/mcp/rightmessage

When prompted, add the authorization header:
- Header: Authorization
- Value: Bearer YOUR_API_TOKEN_HERE

Or manually edit ~/.gemini/settings.json (macOS/Linux) or %USERPROFILE%\.gemini\settings.json (Windows):

{
  "mcpServers": {
    "rightmessage": {
      "httpUrl": "https://api.rightmessage.com/mcp/rightmessage",
      "headers": {
        "Authorization": "Bearer YOUR_API_TOKEN_HERE"
      }
    }
  }
}

Verify with gemini /mcp in your terminal.

What your AI can do

Once connected, your AI assistant can manage nearly every part of your RightMessage account through conversation:

Build and edit segment groups, questions, flows, widgets, campaigns, offers, forms, phrases, goals, and personalized emails
Analyze performance of campaigns, offers, flows, and individual segments, including per-segment lift vs. your baseline
Run split tests and review the results
Manage integrations with your ESP or CRM
Export survey responses and captured leads
Search visitors and review individual visitor profiles and activity
Publish, version, and roll back your configuration

All changes your AI makes are saved as drafts. Nothing goes live until you publish.

For the full tool inventory and the workflow patterns your AI can follow, see rightmessage.com/mcp.

How to work with your AI

Start by asking your AI to show you what projects it can access:

"Show me my RightMessage projects and what's in each one."

Your AI will list your projects with counts of segments, questions, campaigns, and other resources. Tell it which project to work on:

"Set the active project to [your project name]."

Now you can ask it to make changes or analyze data:

"Create a new segment group called 'Experience Level' with three segments: Beginner, Intermediate, and Advanced."
"Show me which campaigns are performing best this month."
"Build a flow that asks visitors about their biggest challenge, then shows them a relevant lead magnet."
"Export all survey responses from the last 30 days."

When you're ready to make changes live, ask your AI to publish:

"Publish these changes to my live site."

Create resources in dependency order for best results: segment groups → questions → offers → flows → widgets → campaigns. Your AI knows this workflow and will follow it automatically.

Troubleshooting

AI can't see RightMessage tools

Verify you fully restarted your AI client (not just closed the window)
Check that your API token is correctly formatted in the config file
Make sure there are no syntax errors in your JSON configuration

Permission denied or authentication errors

Your API token may have been deleted or expired. Generate a new one from your Profile.
Confirm you copied the entire token, including any trailing characters

Rate limit errors

The MCP server allows 120 requests per minute. If you hit this limit, wait 60 seconds and try again.

Changes not appearing on your site

Changes are saved as drafts by default. Ask your AI to run publish-config or manually publish from your RightMessage dashboard.

Was this helpful?