Camoufox Browser

Provides stealthy MCP access to web browsing with persistent sessions and content inspection for AI agents.

python

0

GitHub Stars

python

Language

2 months ago

First Indexed

3 weeks ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "danielmiranda-camoufox-browser-mcp": {
      "command": "docker",
      "args": [
        "exec",
        "-i",
        "camofox-mcp",
        "python",
        "src/mcp_server.py"
      ]
    }
  }
}

Camoufox MCP Server enables AI agents to interact with the modern web through a stealthy, persistent web navigation session powered by the Camoufox engine. It exposes a set of MCP endpoints that let agents navigate, interact, extract content, and capture screenshots while maintaining session continuity across multiple contexts.

How to use

You run this MCP server locally (or in a container) and connect your MCP client to it to perform web interactions. Start a persistent session by navigating to a URL, then reuse that session hash for subsequent actions like clicking, typing, scrolling, or listing links. You can also capture screenshots or retrieve the current page HTML for inspection.

Typical usage pattern:

Create or reuse a session by navigating to a URL.
Interact with the page by clicking elements, typing text, and scrolling.
Extract links or convert the page to Markdown for efficient token usage.
Take a screenshot or fetch the raw HTML for verification.

How to install

Prerequisites you need before installation:

Docker and Docker Compose installed on your host.
Python 3.11+ installed on your host.

Install and start the MCP server via Docker using the following steps:

# Clone and start
git clone https://github.com/danielmiranda/camoufox-browser-mcp.git
cd camoufox-browser-mcp
docker-compose up --build -d

Configuration and running by MCP clients

Configure your MCP client to connect to the Camoufox MCP Server. Use the following JSON configuration snippet in your MCP client configuration file to point to the running MCP server container and execute the server script inside the container.

{
  "mcpServers": {
    "camoufox-browser": {
      "command": "docker",
      "args": ["exec", "-i", "camofox-mcp", "python", "src/mcp_server.py"]
    }
  }
}

Additional notes

For verification, you can run example scripts provided with the project to simulate a full agent flow and to test navigation, link extraction, and interaction sequences.

Available tools

browser_navigate

Navigates to a URL and creates or reuses a persistent session, returning a session hash to continue with that context.

browser_interact

Performs actions such as click, type, scroll_up, and scroll_down within the current session context.

browser_get_markdown

Converts the current page content into Markdown for more token-efficient processing by the AI agent.

browser_list_links

Extracts all links on the current page along with their descriptive text.

browser_screenshot

Captures the current viewport content and returns it as Base64, with an option to save to a file.

browser_snapshot

Retrieves the raw HTML content for the active session.

browser_sessions

Lists all active navigation session hashes.