eyalzh/browser-control-mcp

Overview

A Model Context Protocol server that pairs with a browser extension, allowing AI agents to control the user's browser. It enables programmatic navigation, clicking, form filling, and data extraction through the extension's bridge.

Best for

Best for
Developers who need to give AI agents hands-on browser control for testing or workflow automation

Use cases

• Automate repetitive web tasks like form submissions or data scraping
• Let AI assistants navigate and interact with web apps during development
• Create end-to-end browser tests controlled by language model agents

Notes

A Model Context Protocol server that pairs with a browser extension, allowing AI agents to control the user’s browser. It enables programmatic navigation, clicking, form filling, and data extraction through the extension’s bridge.

287 stars on GitHub. Last updated 2026-03-31. Licensed MIT.

Use cases

• Automate repetitive web tasks like form submissions or data scraping
• Let AI assistants navigate and interact with web apps during development
• Create end-to-end browser tests controlled by language model agents

Pros

• Direct browser control via MCP integrates with many AI agent frameworks
• Open source with a client-side extension keeps user data local
• TypeScript codebase is accessible for custom modifications

Cons

• Requires installing both a browser extension and MCP server
• Limited to scenarios where the extension is actively running and authorized
• May face restrictions on cross-origin or complex JavaScript-heavy sites

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.