Reasonix Browser Bridge
Browser automation for AI agents: tabs, cookies, arbitrary JS execution (via CDP, bypasses CSP), screenshots, downloads, proxy switching, data cleanup. Chrome & Edge multi-browser, full permissions, zero configuration. 25+ MCP tools.
README
Reasonix Browser Bridge
Reasonix 的自动化浏览器插件(开源)—— 通过 MCP 让 Reasonix 直接控制你的浏览器:标签页、Cookie、任意 JavaScript 执行。Chrome / Edge 通用。
┌─────────────┐ stdio / MCP ┌──────────────────┐ WebSocket ┌───────────────┐
│ MCP Client │ ⇄─────────────⇄ │ MCP Server (Node)│ ⇄ 127.0.0.1:8747 ⇄ │ Extension │
│ (Reasonix…) │ └──────────────────┘ │ (MV3) │
└─────────────┘ └───────┬───────┘
│
tabs / cookies /
scripting / debugger (CDP)
▼
Browser
- Extension (
extension/): Manifest V3, connects tows://127.0.0.1:8747, auto-reconnects. Install once per browser (Chrome and Edge). - MCP Server (
server/): stdio MCP server that bridges to the extension and exposes browser tools. - No special flags needed to launch the browser — no
--remote-debugging-port, no native host / registry setup. Install the extension, run the server, done.
Features / Tools
| Tool | Description |
|---|---|
browser_list |
List all connected browsers (chrome / edge) |
browser_state |
Extension connection status, version, heartbeat |
browser_tabs_list |
List all tabs (optional URL filter) |
browser_tabs_open |
Open a new tab |
browser_tabs_activate |
Switch to a tab |
browser_tabs_close |
Close one or more tabs |
browser_tabs_reload |
Reload a tab |
browser_cookies_getAll / get / set / remove |
Full cookie access |
browser_script_execute |
Run arbitrary JavaScript on any page via CDP (bypasses page CSP) |
browser_script_execute_batch |
Run multiple JS snippets in ONE debugger attach (fewer toolbar flashes) |
browser_screenshot |
Screenshot a tab (base64 PNG/JPEG, AI can see the page) |
browser_downloads_download / search |
Trigger / query downloads |
browser_proxy_get / set |
Read / switch proxy settings (direct / system / fixed / PAC) |
browser_browsingdata_remove |
Clear cache / cookies / history / localStorage ... |
browser_management_list |
List all installed extensions (with enabled state) |
browser_history_search |
Search browsing history |
browser_bookmarks_list |
Read bookmark tree |
browser_notify |
System notification |
browser_webnav_frames |
List iframes of a tab |
browser_storage_get / set |
Read/write the extension's local storage |
Auto-update check: the extension checks GitHub Releases every 12h (and on browser start). When a new version is published you'll get a system notification, an orange ↑ badge on the toolbar icon, and a "Download" button in the popup — no store required (works for Developer-mode installs).
Multi-browser: Chrome and Edge are both supported. Install the extension in each browser you want to control — every instance auto-connects to the bridge. Most tools accept an optional browser parameter ("chrome" / "edge") to target a specific browser; when only one browser is connected it is used automatically, and when several are connected without a browser argument the server tells you to specify one.
Install
1. Install the extension (Chrome / Edge — install once per browser)
- Open
chrome://extensions(Chrome) oredge://extensions(Edge). - Enable Developer mode.
- Click Load unpacked / 加载未打包的扩展程序 and select the
extension/folder. - Pin it if you like — the icon shows bridge status (green M = connected).
Note: Chrome 137+ blocks the
--load-extensioncommand-line flag on branded builds; the developer-mode UI install above is the supported way. The same extension folder works in both browsers (same extension ID).
2. Start the MCP server
cd server
npm install
npm start
You should see ✅ 插件已连接 / plugin connected once the extension auto-connects.
3. Configure your MCP client
Add to your MCP client config (Reasonix config.toml):
{
"mcpServers": {
"reasonix-browser-bridge": {
"command": "node",
"args": ["/absolute/path/to/server/index.js"]
}
}
}
Usage example
Ask your AI:
List my open tabs. Then run
document.titleon tab #123.
It maps to browser_tabs_list and browser_script_execute.
browser_script_execute takes a JS expression (no return keyword; use an IIFE for async):
document.title
({ count: document.querySelectorAll('a').length })
(async () => { const r = await fetch('/api/data').then(r => r.json()); return r; })() // set awaitPromise: true
Security
- The WebSocket bridge binds to 127.0.0.1 only; the extension only connects to localhost.
- No remote access, no native messaging, no registry changes.
- The extension requests full permissions (tabs, cookies, scripting, debugger, downloads, proxy, browsingData, management, history, bookmarks, clipboard, notifications, ... +
<all_urls>) — it is designed as Reasonix's automation bridge, so it can do anything the browser can. Install it only if you trust this project.
Project layout
browser-mcp-bridge/
├── extension/ # MV3 browser extension
│ ├── manifest.json
│ ├── background.js # WS bridge + browser gateway
│ ├── popup.* # minimal status popup
│ ├── content.js # page marker (document_start)
│ └── icons/
├── server/ # MCP server (Node)
│ ├── index.js
│ └── package.json
├── README.md
└── LICENSE
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.