gemini-browser-mcp
v1.1.3
Published
MCP server that bridges Claude Code to Google Gemini via a Chrome extension
Maintainers
Readme
gemini-browser-mcp
MCP server that bridges Claude Code to Google Gemini via a Chrome extension — no API key required.
Instead of using the Gemini API, it interacts with the Gemini web app directly, meaning you get all the features available in your browser: Flash, Pro, Extended thinking, uploads, and more.
Quick Start
npx gemini-browser-mcpThis runs the setup wizard, which configures Claude Code for you automatically.
Manual Setup
1. Install the Chrome Extension
Download the extension from the GitHub releases page or clone the repo and load it as an unpacked extension:
git clone https://github.com/Walkoud/gemini-browser-mcp.gitThen load the chrome-extension/ folder in Chrome:
- Go to
chrome://extensions/ - Enable Developer mode (top-right toggle)
- Click Load unpacked and select the
chrome-extension/folder
Make sure you're logged in at gemini.google.com.
2. Configure Claude Code
npx gemini-browser-mcpOr add it manually:
claude mcp add gemini-browser-mcp -- npx gemini-browser-mcp startUsage
In Claude Code, use the ask_gemini_web tool to send prompts to Gemini.
Examples
ask_gemini_web 1+1
ask_gemini_10 web Explain quantum computing in simple terms
ask_gemini_web Write a haiku about MCP serversModel Support
Whatever model you select in the Gemini web UI is what gets used — Flash, Pro, Extended thinking, etc. The extension detects the current mode and adjusts the timeout accordingly (longer for Extended thinking).
The popup also shows the active model and thinking mode at a glance.
How it works
- The Chrome extension connects to a local WebSocket server
- Claude Code sends prompts to Gemini through the web interface (content script)
- The extension waits for the response to appear in the DOM and relays it back
- No API keys needed — everything goes through your browser session
