Agent Browser
Controls a locally-running Chrome browser with preserved login states, supporting CDP commands, screenshots, and keyboard/mouse in
This MCP server enables AI agents to control a real Chrome browser instance while maintaining session state, cookies, and login credentials. It uses a three-layer architecture: an injected Chrome extension, a local TMWebDriver bridge service, and the MCP service layer. Supported operations include page scanning, content extraction, JavaScript execution, Chrome DevTools Protocol commands, screenshots, and keyboard/mouse input. Compatible with Claude Desktop, Cursor, and Hermes Agent.
Source
Repository: https://github.com/335234131/agent-browser-mcp
Maintain Agent Browser?
Let people know it's listed here — add the badge (live metrics, light/dark aware) or a plain link to your README or docs.
[](https://getagentictools.com/mcp/335234131-agent-browser-mcp?ref=badge) Related mcp
OpenBrowser AI-powered browser automation using CodeAgent and Chrome DevTools Protocol.
Mermaid Diagram Generator Converts Mermaid diagrams to PNG images using Puppeteer for high-quality headless browser rendering, supporting multiple themes an
Uber Eats Integrates with Uber Eats using browser automation to enable restaurant searches, menu exploration, and food order placement for a
HT Terminal Provides direct access to headless terminal sessions through a high-performance Rust implementation, enabling terminal automation,
Google Search Automates parallel Google searches with intelligent CAPTCHA detection, browser state persistence, and user behavior simulation to
Google Surf Search Google and extract article content using Playwright with a persistent Chrome profile, no API key required.