MCP Browser Agent
An MCP (Model Context Protocol) server that enables AI assistants to automate browser-based tasks, extract structured web data, download files, and organize local outputs using Python.
The project integrates FastMCP, browser-use, and Playwright to provide browser automation tools to compatible MCP clients such as Cursor and Claude Desktop.
⚠️ Project status: In active development. Features and setup instructions may evolve as the implementation matures.
Features
- MCP Integration: Exposes browser automation capabilities as tools that compatible AI clients can discover and invoke.
- AI-Powered Browsing: Uses an autonomous browser agent to navigate complex websites and perform multi-step tasks.
- Structured Data Extraction: Collects web information and returns clean JSON results.
- File Downloads & Organization: Saves and automatically categorizes downloaded files into structured local directories.
- Python-Based Architecture: Modular setup keeping server, browser automation, configuration, and file management separated.
Architecture
text
┌──────────────────────────────┐
│ MCP Client │
│ Cursor / Claude Desktop │
└──────────────┬───────────────┘
│
│ MCP over stdio
▼
┌──────────────────────────────┐
│ FastMCP Server │
│ │
│ Browser tools │
│ File management tools │
│ Configuration │
└──────────────┬───────────────┘
│
┌───────┴────────┐
▼ ▼
┌──────────────┐ ┌────────────────┐
│ Browser Agent│ │ File Manager │
│ browser-use │ │ pathlib/shutil │
│ Playwright │ │ │
└──────┬───────┘ └───────┬────────┘
│ │
▼ ▼
┌──────────────┐ ┌────────────────┐
│ Websites │ │ Local Workspace│
│ and Web Data │ │ Downloads/JSON │
└──────────────┘ └────────────────┘
Top comments (0)