
MCP Browser
Connects your AI assistant to a real Chromium browser session (Chrome, Edge, or Brave) for web automation that bypasses anti-bot protections, handles authentication, and works with JavaScript-heavy sites. Exposes eight browser operations: fetch pages, execute JavaScript, click elements, type text into forms, extract current HTML, scroll, screenshot, and close tabs. Built on Puppeteer but adds an intelligence layer that auto-detects SPAs (React, Vue, Angular), follows SSO redirects, discovers form fields with constraints, and returns cleaned HTML that's 90% smaller than raw DOM. Reuses tabs per domain and lets you extract updated content without reloading. Useful when your agent needs to interact with authenticated dashboards, corporate intranets, or any site that blocks headless browsers.
- GitHub stars
- 9
- tools
- 8
- transport
- STDIO
- last updated
- Nov 8, 2026
Tools
8 toolsPublic tool metadata for what this MCP can expose to an agent.
browser_fetch_webpageFetches web pages using your Chrome/Edge/Brave browser. Handles authentication, CAPTCHA, SSO, anti-bot protection, and JavaScript-heavy sites.
Fetches web pages using your Chrome/Edge/Brave browser. Handles authentication, CAPTCHA, SSO, anti-bot protection, and JavaScript-heavy sites.
No parameter schema in public metadata yet.
browser_execute_javascriptExecutes a JavaScript snippet in the active page context and returns the result with metadata (execution time, truncation flag, URL-change detection).
Executes a JavaScript snippet in the active page context and returns the result with metadata (execution time, truncation flag, URL-change detection).
No parameter schema in public metadata yet.
browser_click_elementClicks on any clickable element (buttons, links, divs with onclick handlers, etc.). Can target by CSS selector or visible text content.
Clicks on any clickable element (buttons, links, divs with onclick handlers, etc.). Can target by CSS selector or visible text content.
No parameter schema in public metadata yet.
browser_type_textTypes text into one or more input fields in a single call. Supports filling entire forms at once for efficient automation.
Types text into one or more input fields in a single call. Supports filling entire forms at once for efficient automation.
No parameter schema in public metadata yet.
browser_get_current_htmlGets the current HTML from an already-loaded page WITHOUT navigating or reloading.
Gets the current HTML from an already-loaded page WITHOUT navigating or reloading.
No parameter schema in public metadata yet.
browser_scroll_pageScrolls within an already-loaded page, by direction, to an element, or to an absolute position.
Scrolls within an already-loaded page, by direction, to an element, or to an absolute position.
No parameter schema in public metadata yet.
browser_take_screenshotTakes a screenshot of an already-loaded page for visual analysis and returns a PNG image.
Takes a screenshot of an already-loaded page for visual analysis and returns a PNG image.
No parameter schema in public metadata yet.
browser_close_tabCloses the browser tab for the given URL's hostname and forces a fresh session on the next visit to that hostname.
Closes the browser tab for the given URL's hostname and forces a fresh session on the next visit to that hostname.
No parameter schema in public metadata yet.
Sponsors
Readme
✅ MCPBrowser (MCP Browser)
Privacy: Pages you browse are shared with your AI assistant and its model provider. Only access content you're authorized to share. See privacy and security details.
Details
- Repository
- github.com/cherchyk/mcpbrowser
- Website
- github.com/cherchyk/MCPBrowser#readme
- Categories
- Web & Browser Automation, Automation & Workflows
- Package
mcpbrowser- Transport
- STDIO
- Registry
- MCP registry, active