MCP Browser

MCP Browser avatar

by neoforge-dev

Browser Automation

NeoForge Browser MCP server - used to test the frontend

What is MCP Browser

MCP Browser

A headless browser interface for the Model Control Protocol (MCP).

Features

  • Headless browser automation using Playwright
  • Web UI for browser interaction
  • WebSocket communication for real-time updates
  • Real-time browser event subscription system
  • Integration with MCP for AI agents

Prerequisites

  • Python 3.13+
  • uv for dependency management
  • Docker (for containerized usage)

Local Development

Setup with uv

# Clone the repository
git clone https://github.com/yourusername/mcp-browser.git
cd mcp-browser

# Install dependencies
uv venv .venv
source .venv/bin/activate
uv pip install -e .

# Install Playwright browsers
python -m playwright install

Running Locally

For a simple test without Xvfb:

./simple_test.sh

For a full test with Xvfb (requires X11):

./test_local.sh

Docker Deployment

Build and run using Docker Compose:

# Set your MCP secret
export MCP_SECRET=your_secret_key

# Build and run
docker-compose up --build

Or use the provided script:

./run.sh

Configuration

The following environment variables can be set:

  • MCP_SECRET: Secret key for MCP authentication
  • SERVER_PORT: Port to run the server on (default: 7665)
  • PLAYWRIGHT_SKIP_BROWSER_DOWNLOAD: Set to 1 to skip browser download and run in headless-only mode

API Endpoints

  • GET /: Web UI
  • GET /api/status: Get browser and MCP client status
  • WebSocket /ws: WebSocket endpoint for real-time communication
  • WebSocket /ws/browser/events: WebSocket endpoint for browser event subscriptions
  • GET /api/browser/subscribe: Subscribe to browser events
  • GET /api/browser/unsubscribe: Unsubscribe from browser events
  • GET /api/browser/subscriptions: List active event subscriptions

Event Subscriptions

The MCP Browser supports real-time event subscriptions via WebSockets. This allows clients to receive browser events as they happen, including:

  • Page events (navigation, load, error)
  • DOM events (mutations, changes)
  • Console events (logs, warnings, errors)
  • Network events (requests, responses, errors)

For detailed documentation and examples of the event subscription system, see:

License

MIT

Leave a Comment

Comments section will be available soon. Stay tuned!

Frequently Asked Questions

What is MCP?

MCP (Model Context Protocol) is an open protocol that standardizes how applications provide context to LLMs. Think of MCP like a USB-C port for AI applications, providing a standardized way to connect AI models to different data sources and tools.

What are MCP Servers?

MCP Servers are lightweight programs that expose specific capabilities through the standardized Model Context Protocol. They act as bridges between LLMs like Claude and various data sources or services, allowing secure access to files, databases, APIs, and other resources.

How do MCP Servers work?

MCP Servers follow a client-server architecture where a host application (like Claude Desktop) connects to multiple servers. Each server provides specific functionality through standardized endpoints and protocols, enabling Claude to access data and perform actions through the standardized protocol.

Are MCP Servers secure?

Yes, MCP Servers are designed with security in mind. They run locally with explicit configuration and permissions, require user approval for actions, and include built-in security features to prevent unauthorized access and ensure data privacy.