How do I use chrome-pilot-mcp?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@chrome-pilot-mcp Navigate to google.com and search for 'MCP protocol'" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

chrome-pilot-mcp

by HelloiOS2014

Overview Schema Source Related Servers Discussions Admin

TypeScript

Local

chrome-pilot-mcp

English

通过 CDP（Chrome DevTools Protocol）连接已运行的 Chrome，并以 MCP 服务的形式暴露浏览器控制能力。

项目基于 patchright-core，适合需要保留登录态、Cookie、浏览器扩展和现有标签页的自动化场景。

核心特点

连接已有 Chrome 实例，无需重新启动浏览器。
保留登录态、用户数据和当前标签页上下文。
基于 patchright-core，尽量降低常见 CDP 自动化指纹。
提供连接、标签页、导航、交互、检视、表单等 18 个 MCP 工具。

Related MCP server: Real Browser MCP

安装

Claude Code（推荐）

# 全局安装
claude mcp add --scope user chrome-pilot -- npx -y chrome-pilot-mcp

# 项目级安装
claude mcp add --scope project chrome-pilot -- npx -y chrome-pilot-mcp

--scope user 写入 ~/.claude/settings.json，--scope project 写入 .claude/settings.local.json。

通过 npx 直接运行

npx -y chrome-pilot-mcp

手动配置

{
  "mcpServers": {
    "chrome-pilot": {
      "command": "npx",
      "args": ["-y", "chrome-pilot-mcp"]
    }
  }
}

从源码运行

git clone <repo-url>
cd chrome-pilot-mcp
npm install
npm run build
npm start

前置条件

Node.js / npx

npx 用于 Claude Code 启动 MCP 子进程。请先确认 Node.js 与 npx 可用（Node >= 18）：

node -v
npx -v

如果缺失可按平台安装 Node.js（安装后会包含 npx）：

# macOS
brew install node

# Ubuntu / Debian
sudo apt-get update && sudo apt-get install -y nodejs npm

# Windows (PowerShell)
winget install OpenJS.NodeJS.LTS

Chrome 版本

需要 Chrome >= 144，以支持通过 chrome://inspect/#remote-debugging 启用远程调试。

启用远程调试

在 Chrome 地址栏打开 chrome://inspect/#remote-debugging 并启用远程调试。Chrome 重启后通常需要重新启用。

DevToolsActivePort 路径

平台	路径
macOS	`~/Library/Application Support/Google/Chrome/DevToolsActivePort`
Linux	`~/.config/google-chrome/DevToolsActivePort`
Windows	`%LOCALAPPDATA%\\Google\\Chrome\\User Data\\DevToolsActivePort`

工具概览

类别	工具	用途
连接	`chrome_status`, `chrome_connect`, `chrome_disconnect`	管理 Chrome 连接状态
标签页	`chrome_list_tabs`, `chrome_select_tab`, `chrome_close_tab`	查看并切换标签页
导航	`chrome_navigate`, `chrome_back`, `chrome_forward`, `chrome_reload`	执行页面导航
交互	`chrome_click`, `chrome_type`, `chrome_scroll`, `chrome_press_key`	与页面元素交互
检视	`chrome_screenshot`, `chrome_dump_dom`, `chrome_evaluate`	截图并检查页面状态
表单	`chrome_fill_form`, `chrome_select_option`	批量填写表单

详细文档

中文使用手册：docs/usage.md
English usage guide: docs/usage.en.md

License

MIT

Available Tools

19 tools

chrome_backA

Go back in browser history

Inputs: tab_index.

chrome_clickB

Click on an element by CSS selector or coordinates

Inputs: x, y, selector, tab_index, +1 more.

chrome_close_tabB

Close a tab by index

Inputs: index (required).

chrome_connectA

Connect to an already-running Chrome browser via DevTools Protocol. Chrome must have remote debugging enabled (chrome://inspect/#remote-debugging).

Inputs: no named parameters.

chrome_disconnectA

Disconnect from Chrome (does not close the browser)

Inputs: no named parameters.

chrome_dump_domA

Get the DOM HTML of the page or a specific element

Inputs: outer, selector, tab_index.

chrome_evaluateB

Execute JavaScript in the page context and return the result

Inputs: expression (required), tab_index.

chrome_fill_formB

Fill multiple form fields at once

Inputs: fields (required), tab_index, clear_first.

chrome_forwardB

Go forward in browser history

Inputs: tab_index.

chrome_list_tabsA

List all open tabs with their title and URL

Inputs: no named parameters.

chrome_navigateA

Navigate the active tab to a URL

Inputs: url (required), tab_index, wait_until.

chrome_press_keyC

Press a keyboard key

Inputs: key (required), modifiers, tab_index.

chrome_reloadB

Reload the current page

Inputs: tab_index, ignore_cache.

chrome_screenshotA

Take a screenshot of the current page (returns base64 image)

Inputs: format, quality, selector, full_page, +1 more.

chrome_scrollC

Scroll the page or a specific element

Inputs: amount, selector, direction, tab_index.

chrome_select_optionA

Select an option from a dropdown

Inputs: selector (required), label, value, tab_index.

chrome_select_tabA

Switch to a tab by index (from chrome_list_tabs)

Inputs: index (required).

chrome_statusB

Check Chrome connection status

Inputs: no named parameters.

chrome_typeC

Type text into a focused element or a specific element by selector

Inputs: text (required), delay, selector, tab_index, +1 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

Deploy Server

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A3.5/5.0

Scored 2026-08-09 21:05 across 19 tools

Disambiguation5/5

Each tool has a clear, dedicated purpose: lifecycle (connect/disconnect/status), tab management (list/select/close), navigation (navigate/back/forward/reload), interaction (click/type/scroll/press_key), page inspection (screenshot/dump_dom/evaluate), and form helpers (fill_form/select_option). No two tools target the same action; even type and fill_form are distinguished by single vs. multi-field.

Naming Consistency5/5

All tools are lowercase snake_case with a uniform chrome_ prefix. The second part nearly always follows a verb or verb_noun pattern (e.g., list_tabs, select_tab, click, navigate). This makes the tool set predictable and easy to navigate.

Tool Count3/5

At 19 tools, the server is on the heavier side. While each tool is justifiable, several shortcuts (fill_form, select_option) could be composed from more primitive tools, and the count feels more like a comprehensive library than a minimal set. This is borderline appropriate for a browser automation tool.

Completeness4/5

Core browser operations are covered: tab lifecycle, navigation, input, screenshots, DOM access, and JS evaluation. However, there is no direct way to create a new tab or wait for conditions, which falls short of full browser automation coverage. These gaps are workable via evaluate, so minor.

Maintenance

ActivityInactive

ResponsivenessNo recent community issues

Resources

Related MCP Connectors

Tap
io.github.LeonTing1010
Browser MCP for logged-in tasks. Uses your Chrome — credentials stay local. Zero-token replay.
Browser Automation Web Scraping
218
https://mcp.taprun.dev/mcp
browser-mcp
io.github.Shreyas-Profile
Hosted real Google Chrome MCP with per-user persistent state. Navigate, click, type, screenshot.
Browser Automation Web Scraping
2
https://browser.regiq.in/api/mcp
BrowserlessA
io.github.pipeworx-io
Browserless MCP — wraps the Browserless headless-Chromium REST API (browserless.io)
Browser Automation Web Scraping
481
https://gateway.pipeworx.io/browserless/mcp
BrowzA
com.gobrowz.mcp
Browz hosted MCP (free public beta): let an AI app drive your own Browz browser profiles from anywhere. Connect https://mcp.gobrowz.com/mcp and sign in with your Browz account (OAuth) or an agent token. Each call is relayed to the Browz app on your computer, which must be online with Remote control on and have a hardware security chip (Mac Secure Enclave or TPM 2.0). 17 tools: profiles, tabs, navigate, snapshot, act, wait, screenshot. Cookies, downloads and stored sign-ins stay on your computer.
Browser Automation
171
https://mcp.gobrowz.com/mcp

Related MCP Servers

playwriter
Browser Automation Web Scraping
remorses
A
license
Not graded
quality
A
maintenance
Connects to your running Chrome browser, enabling AI agents to automate it via Playwright API with existing logins and extensions.
Updated 2 days ago (2026-11-12 21:23 UTC)
4,045 npm
4,080
MIT
Real Browser MCP
Browser Automation Web Scraping
ton-to-ton
A
license
Not graded
quality
D
maintenance
Enables browser automation over MCP using a real Chrome browser with existing profile, supporting real tabs, downloads, cookies, and RPA workflows.
Updated 5 months ago (2026-06-19 04:37 UTC)
166 npm
MIT
chrome-remote-debugging-mcp
Browser Automation Web Scraping
danilin-em
A
license
A
quality
B
maintenance
Controls a running Chrome/Chromium browser via the Chrome DevTools Protocol, enabling navigation, JavaScript evaluation, tab management, and raw CDP commands through MCP tools.
Updated 24 days ago (2026-10-21 18:11 UTC)
5
MIT
Chrome Live MCP
Browser Automation Agent Orchestration
wankaiyi
A
license
Not graded
quality
B
maintenance
Connects local stdio MCP servers to an existing Chrome 144+ session, preserving the user's signed-in sessions, cookies, tabs, and extension environment without launching a second browser. It provides tab control, semantic snapshots, screenshots, pointer, keyboard, form selection, scrolling, navigation, and waiting tools.
Updated 3 months ago (2026-08-24 13:24 UTC)
MIT

chrome-pilot-mcp

核心特点

安装

Claude Code（推荐）

通过 npx 直接运行

手动配置

从源码运行

前置条件

Node.js / npx

Chrome 版本

启用远程调试

DevToolsActivePort 路径

工具概览

详细文档

License

Available Tools

Tool Schema Changelog

TDQS

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

playwriter

Real Browser MCP

chrome-remote-debugging-mcp

Chrome Live MCP