open-claude-in-chrome MCP server

Claude in Chrome, reverse-engineered and open-source. No domain blocklist. Any Chromium browser. Same 18 MCP tools, same performance.

What open-claude-in-chrome does

Summarised from the project README; check the repository for the latest details.

Tools exposed (20+)

• tabs_context_mcp – Get tab group context
• tabs_create_mcp – Create a new tab
• tabs_close_mcp – Close a tab
• navigate – Navigate to URL, back, forward
• computer – Mouse, keyboard, screenshot, zoom
• read_page – Accessibility tree with element refs
• get_page_text – Extract article/main text
• find – Find elements by text/attributes
• form_input – Set form values by ref
• javascript_tool – Execute JS in page context
• read_console_messages – Console output (filtered)
• read_network_requests – Network activity
• resize_window – Resize browser window
• file_upload – Attach local file(s) to a file input (by ref)
• upload_image – Attach a captured screenshot to a file input (by ref)
• gif_creator – GIF recording
• shortcuts_list – List shortcuts
• shortcuts_execute – Run a shortcut
• switch_browser – Hand off automation to another Chromium browser
• execute_code – Run sandboxed JS that drives every tool via `chrome.`

Requirements

• Node.js v18+
• Any Chromium browser (Chrome, Edge, Brave, Arc, Opera, Vivaldi, etc.)
• Claude Code v2.1.80+ (the recorder needs channels; browser automation alone works on v2.0.73+)
• An OpenAI API key (used to transcribe recording narration)

How to install open-claude-in-chrome

Listed on Socketlist – add this badge to your README:

Badge markdown[](https://socketlist.pages.dev/s/noemica-io-open-claude-in-chrome/)

Maintain this project? Get it featured for free during launch →

Compare with chrome-devtools-mcp →

Related Browser Automation MCP servers