MCPMonkey

by kstrikis

331 downloads
Not rated
GitHub

About

MCPMonkey is an MCP server and Firefox extension so your AI agent can access your browser

Details

Author
kstrikis
Downloads
331
Categories
Automation

- Tab management: get info, create, close, activate, and duplicate tabs.
- Page style extraction: extract global styles, computed styles, color schemes, typography.
- Full compatibility with existing Violentmonkey userscripts.
- Planned MCP server management and installation from the browser.
- Planned fine-grained permissions control over MCP server resources.
- Uses WebSocket (port 3025) internally and stdio transport for external MCP clients.

Setting up with Highlight

This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:

  1. Download and install Highlight from highlightai.com/download
  2. Navigate to the plugins tab and select "Add Custom Plugin"
  3. Configure the plugin with the settings below
    Plugin Name MCPMonkey
    Command (node, npx, python, etc.)

    Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.

  4. Enable "Start Automatically" if you want the plugin to start when Highlight launches

From the repository

Install two components: the MCP server and the browser extension. In Cursor, add the MCP server via Settings > Features > MCP Servers with the command npx mcpmonkey-server. Install the browser extension by building from source and loading it as a temporary add-on in Firefox (.xpi files are planned).

Claude Desktop / Cursor

Paste into your MCP client config file to install this server.

{
    "mcpServers": {
        "mcpmonkey": {
            "MCPMonkey": {
                "command": "node",
                "args": [
                    "build/index.js"
                ]
            }
        }
    }
}

McpServers

{
    "MCPMonkey": {
        "command": "node",
        "args": [
            "build/index.js"
        ]
    }
}

MCPMonkey

MCPMonkey is a fork of Violentmonkey, extending its powerful userscript capabilities to support Model Context Protocol (MCP) servers. This project aims to bridge the gap between AI language models and browser interactions.

About MCPMonkey

MCPMonkey enhances the browser extension capabilities of Violentmonkey to provide a user-friendly interface for managing and using Model Context Protocol (MCP) servers. This allows AI language models like those used in Cursor to interact with your browser in meaningful ways.

Current Features

MCPMonkey currently implements these key features:

- Tab Management (browserAction Tool):
- Get information about all open tabs
- Create new tabs
- Close tabs
- Activate (focus) tabs
- Duplicate tabs

- Page Style Extraction (getPageStyles Tool):
- Extract styling information from web pages for AI analysis
- Returns structured JSON data about styles
- Includes global styles, computed styles, color schemes, typography

- User Script Support: Full compatibility with existing userscripts (inherited from Violentmonkey)

Planned Features

The following features are planned for future development:

- MCP Server Management: Install and manage multiple MCP servers directly from your browser
- Enhanced Browser Access: Allow AI tools to interact with:
- Browsing history
- Bookmarks
- Dev console logs
- Other installed extensions
- Permissions Control: Fine-grained control over what resources each MCP server can access
- Community Hub: Share and discover MCP servers and scripts

Cursor MCP Integration

MCPMonkey consists of two main components that need to be installed in the following order:
1. MCP Server
2. Browser Extension

1. Adding the MCP Server to Cursor

Using the published NPM package:

- In Cursor, go to Settings > Features > MCP Servers
- Click the + Add New MCP Server button
- Configure the connection with:
- Name: mcpmonkey-server
- Type: command
- Command: npx mcpmonkey-server

Or if building yourself:

```sh

No reviews yet — be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.