mcp-browser-use
mcp-browser-use is an MCP server that gives AI assistants like Claude the ability to control a real web browser through automated tasks. It wraps the browser-use library and exposes tools for navigating pages, filling forms, clicking buttons, extracting data, and conducting multi-step web research. The server runs as an HTTP daemon to handle long-running browser automation tasks (30-120+ seconds) without timeout issues, and includes an experimental skills system that learns API endpoints from browser sessions to speed up repeated tasks by 50x.
Key Features
Use Cases
- 01Automating web research by having Claude search multiple sources and compile synthesized reports with key findings
- 02Extracting structured data from websites by instructing the AI to navigate and scrape product prices, listings, or content
- 03Testing web applications through AI-driven browser interactions that fill forms, click buttons, and verify outcomes
- 04Learning reusable API shortcuts from one-time browser sessions to speed up repeated tasks like searching npm packages
- 05Monitoring website changes by periodically running browser tasks that check for specific content or availability
- 06Automating form submissions and multi-step workflows that require navigating complex user interfaces
Related MCP Servers
View moremarkitdown
Python tool for converting files and office documents to Markdown.
firecrawl
The context API to search, scrape, and interact with the web at scale. 🔥
prompts.chat
f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.
langflow
Langflow is a powerful tool for building and deploying AI-powered agents and workflows.
mcp-browser-use — FAQ
What is mcp-browser-use?+
mcp-browser-use is an MCP server that wraps the browser-use library to give AI assistants like Claude the ability to automate real web browsers. It exposes tools for navigation, data extraction, form filling, and multi-step research tasks.
How do I install mcp-browser-use for Claude Desktop?+
Clone the repository, run 'uv sync' and 'uv run playwright install chromium', then start the server with 'uv run mcp-server-browser-use server'. Add the HTTP endpoint (http://localhost:8383/mcp) to your Claude Desktop config using the 'streamable-http' transport type.
Which AI clients work with mcp-browser-use?+
Any MCP client that supports HTTP transport works directly with mcp-browser-use. For clients that only support stdio, you can use mcp-remote as a proxy to connect to the HTTP daemon.
Do I need an API key to use mcp-browser-use?+
Yes, the browser agent requires an LLM API key to decide actions during automation. Set GEMINI_API_KEY, ANTHROPIC_API_KEY, OPENAI_API_KEY, or another supported provider's key as an environment variable or via config.
Is mcp-browser-use free and open source?+
Yes, mcp-browser-use is released under the MIT license and is completely free to use. However, you'll incur costs from your chosen LLM provider (Google Gemini, OpenAI, Anthropic, etc.) for the API calls made by the browser agent.
What is the skills system and should I enable it?+
The skills system is an experimental feature (disabled by default) that learns API endpoints from browser sessions and replays them directly for 50x faster execution. It's in alpha and may have rough edges, but can dramatically speed up repeated tasks once a skill is learned.
How do I install mcp-browser-use?+
Open the source repository on GitHub and follow its README. mcp-browser-use is a mcp server — MCP Agents Market links you directly to the official repo.
Is mcp-browser-use free?+
mcp-browser-use is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.