chrome-devtools-mcp-wrapper
A wrapper around chrome-devtools-mcp that adds dynamic multi-browser routing, allowing connection to and switching between multiple Chrome instances at runtime.
README
chrome-devtools-mcp-wrapper
A dynamic multi-browser wrapper around chrome-devtools-mcp that lets your AI coding agent connect to multiple Chrome instances at once and switch between them at runtime — without restarting the MCP server.
Implements the feature proposed in ChromeDevTools/chrome-devtools-mcp#590.
Key features
- Dynamic browser routing: connect to any number of Chrome instances and route each tool call to the right one via an optional
browserUrlargument. - Runtime switching: call
switch_browserto change the active browser without restarting anything. - All upstream tools included: every tool from
chrome-devtools-mcpis proxied automatically, plus two wrapper-native tools (switch_browser,list_browsers). - No fork: runs
npx chrome-devtools-mcp@latestas a subprocess, so upstream updates are picked up automatically.
Getting started
Add the following config to your MCP client:
{
"mcpServers": {
"chrome-devtools-wrapper": {
"command": "npx",
"args": ["-y", "chrome-devtools-mcp-wrapper@latest"]
}
}
}
To connect to an already-running Chrome instance at startup:
{
"mcpServers": {
"chrome-devtools-wrapper": {
"command": "npx",
"args": ["-y", "chrome-devtools-mcp-wrapper@latest", "--browser-url=http://127.0.0.1:9222"]
}
}
}
[!NOTE] Using
chrome-devtools-mcp-wrapper@latestensures your MCP client always uses the latest version.
MCP client configuration
<details> <summary>Claude Code</summary>
Install via CLI (MCP only)
Use the Claude Code CLI to add the wrapper (guide):
claude mcp add chrome-devtools-wrapper --scope user -- npx -y chrome-devtools-mcp-wrapper@latest
</details>
<details> <summary>Copilot / VS Code</summary>
Click the button to install:
Or install manually:
Follow the VS Code MCP configuration guide and use the standard config above, or use the CLI:
For macOS and Linux:
code --add-mcp '{"name":"chrome-devtools-wrapper","command":"npx","args":["-y","chrome-devtools-mcp-wrapper@latest"],"env":{}}'
For Windows (PowerShell):
code --add-mcp '{"""name""":"""chrome-devtools-wrapper""","""command""":"""npx""","""args""":["""-y""","""chrome-devtools-mcp-wrapper@latest"""]}'
</details>
<details> <summary>Cursor</summary>
Click the button to install:
<img src="https://cursor.com/deeplink/mcp-install-dark.svg" alt="Install in Cursor">
Or install manually:
Go to Cursor Settings → MCP → New MCP Server and paste the standard config above.
</details>
<details> <summary>Kiro</summary>
In Kiro Settings, go to Configure MCP → Open Workspace or User MCP Config and paste the standard config above.
Or from the IDE Activity Bar → Kiro → MCP Servers → Open MCP Config.
</details>
<details> <summary>JetBrains AI Assistant & Junie</summary>
Go to Settings | Tools | AI Assistant | Model Context Protocol (MCP) → Add. Use the standard config above.
For Junie: Settings | Tools | Junie | MCP Settings → Add.
</details>
<details> <summary>Claude Desktop</summary>
Config file locations:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Paste the standard config above.
</details>
<details> <summary>Cline</summary>
Follow https://docs.cline.bot/mcp/configuring-mcp-servers and use the standard config above.
</details>
<details> <summary>Gemini CLI</summary>
gemini mcp add chrome-devtools-wrapper npx chrome-devtools-mcp-wrapper@latest
Or globally:
gemini mcp add -s user chrome-devtools-wrapper npx chrome-devtools-mcp-wrapper@latest
</details>
<details> <summary>Codex</summary>
Follow the configure MCP guide and use the standard config above.
On Windows 11, configure the Chrome install location and increase the startup timeout by updating .codex/config.toml:
[mcp_servers.chrome-devtools-wrapper]
command = "cmd"
args = ["/c", "npx", "-y", "chrome-devtools-mcp-wrapper@latest"]
env = { SystemRoot="C:\\Windows", PROGRAMFILES="C:\\Program Files" }
startup_timeout_ms = 20_000
</details>
Requirements
Usage
1. Start Chrome with remote debugging
macOS
/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome \
--remote-debugging-port=9222 \
--user-data-dir=/tmp/chrome-profile
Linux
google-chrome --remote-debugging-port=9222 --user-data-dir=/tmp/chrome-profile
Windows
"C:\Program Files\Google\Chrome\Application\chrome.exe" --remote-debugging-port=9222 --user-data-dir="%TEMP%\chrome-profile"
[!WARNING] Enabling the remote debugging port exposes a debugging interface that any local application can connect to. Avoid browsing sensitive websites while the port is open.
2. Connect to your browser
switch_browser(url: "9222")
Or with a full URL or WebSocket endpoint:
switch_browser(url: "http://127.0.0.1:9222")
switch_browser(url: "ws://127.0.0.1:9222/devtools/browser/abc123")
After connecting, all chrome-devtools-mcp tools become available.
3. Use any browser tool
navigate_page(url: "https://example.com")
take_screenshot()
evaluate_script(expression: "document.title")
4. Route a single call to a specific browser
Every proxied tool accepts an optional browserUrl argument:
navigate_page(url: "https://staging.example.com", browserUrl: "9333")
take_screenshot(browserUrl: "9222")
5. Check active connections
list_browsers()
Wrapper-native tools
These two tools are always available, even before a browser connects:
| Tool | Description |
|---|---|
switch_browser(url) |
Connect to a browser and set it as the default. Accepts a full URL, WebSocket endpoint, or bare port number (e.g. "9222"). |
list_browsers() |
List all active browser backends and which is the current default. |
All other tools come from chrome-devtools-mcp and are registered once a browser connects.
Configuration options
WebSocket endpoint
If you know the exact WebSocket endpoint (from http://127.0.0.1:9222/json/version → webSocketDebuggerUrl):
{
"mcpServers": {
"chrome-devtools-wrapper": {
"command": "npx",
"args": [
"-y",
"chrome-devtools-mcp-wrapper@latest",
"--browser-url=ws://127.0.0.1:9222/devtools/browser/<id>"
]
}
}
}
Multiple browsers
Start multiple Chrome instances on different ports, then route calls dynamically:
switch_browser(url: "9222")
navigate_page(url: "https://prod.example.com")
navigate_page(url: "https://staging.example.com", browserUrl: "9333")
take_screenshot(browserUrl: "9333")
How it works
MCP client (Claude Code, Cursor, etc.)
│
▼
chrome-devtools-mcp-wrapper
├── switch_browser / list_browsers ← always available
├── navigate_page(browserUrl?) ─┐
├── take_screenshot(browserUrl?) ├─ proxied from chrome-devtools-mcp
├── evaluate_script(browserUrl?) │ registered after first browser connects
└── ... (50+ tools) ─┘
│ │
▼ ▼
backend [9222] backend [9333]
npx chrome- npx chrome-
devtools-mcp devtools-mcp
--browserUrl= --browserUrl=
...9222 ...9333
- One
chrome-devtools-mcpsubprocess per browser URL, reused across calls - Idle backends are killed after 5 minutes of inactivity
- Upstream updates are inherited automatically on restart
Relation to upstream
This package wraps chrome-devtools-mcp without forking it. All 50+ upstream tools are proxied transparently.
For single-browser use cases, the upstream package is the right choice. Use this wrapper when you need to connect to multiple Chrome instances or switch between them at runtime.
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.