Chrome DevTools for Agents
Vendor-official Chrome DevTools MCP server from the Chrome DevTools team at Google. Lets agents control and inspect a live Chrome browser for automation, debugging, and performance analysis.
“Chrome DevTools for Agents gives an AI coding agent the same browser inspection toolkit a human developer uses. It records performance traces and extracts actionable insights, captures screenshots, reads console messages with source-mapped stack traces, and analyses network requests, all driven through Puppeteer for reliable automation. Google ships this as `chrome-devtools-mcp` on npm with full vendor backing from the Chrome DevTools team. The architectural decision worth noting: Google chose to integrate with Chrome DevTools' existing performance API including the Chrome User Experience Report (CrUX) for real-user data, which puts field measurements alongside lab measurements in the agent's hands. The disclosure operators should read carefully: usage statistics are enabled by default, and the README documents how to opt out (`--no-usage-statistics` flag or environment variables). For privacy-sensitive deployments, opting out is one config line.”
INSTALL THIS SERVER
{
"mcpServers": {
"chrome-devtools": {
"command": "npx",
"args": [
"-y",
"chrome-devtools-mcp@latest"
]
}
}
}
{
"mcpServers": {
"chrome-devtools": {
"command": "npx",
"args": [
"-y",
"chrome-devtools-mcp@latest"
]
}
}
}
{
"mcpServers": {
"chrome-devtools": {
"command": "npx",
"args": [
"-y",
"chrome-devtools-mcp@latest"
]
}
}
}
{
"mcpServers": {
"chrome-devtools": {
"command": "npx",
"args": [
"-y",
"chrome-devtools-mcp@latest"
]
}
}
}
{
"mcpServers": {
"chrome-devtools": {
"command": "npx",
"args": [
"-y",
"chrome-devtools-mcp@latest"
]
}
}
}
6 TOOLS AVAILABLE
OUR ASSESSMENT
- Vendor-official from the Chrome DevTools team at Google; same engineers who build Chrome DevTools maintain the MCP server.
- Performance trace recording with CrUX integration places real-user data alongside lab measurements in the agent's context.
- Slim mode (`--slim`) for basic browser tasks reduces tool surface area for agents that only need a handful of capabilities.
- Active development with current Chrome support; commits to the latest Extended Stable Chrome version.
- Officially supports Google Chrome and Chrome for Testing only; other Chromium browsers may work with caveats.
- Usage statistics are enabled by default; operators handling privacy-sensitive workflows should opt out via `--no-usage-statistics` or `CHROME_DEVTOOLS_MCP_NO_USAGE_STATISTICS` env var.
- Browser inspection exposes page content to the MCP client; the README warns against sharing sensitive or personal information that should stay private.
The server exposes browser content to the MCP client, allowing inspection, debugging, and modification of any data in the browser or DevTools. The README warns explicitly against sharing sensitive or personal information through this surface. Usage statistics are collected by Google by default for product improvement, with `--no-usage-statistics` and the `CHROME_DEVTOOLS_MCP_NO_USAGE_STATISTICS` env var providing opt-out paths. Performance trace URLs may be sent to the Google CrUX API for real-user data; `--no-performance-crux` disables this. For CI environments, the `CI` env var also disables usage statistics automatically.
Teams whose agents need to inspect, debug, or measure web applications. Performance tuning, accessibility audits, and visual regression workflows.
TECHNICAL DETAILS
ADOPTION METRICS
// Reading this37.5k stars in seven months reflect strong adoption from agent-driven web development teams. Vendor-official status from the Chrome DevTools team validates the integration depth.
// Reading thisSecond-ranked overall and second in dev-tools, behind OpenSpec. The Puppeteer-backed automation pairs well with web testing workflows.
SOURCES & VERIFICATION
We don't take any single directory's word for it. Before scoring, we cross-reference 5 public MCP sources, install the server ourselves against the clients we cover, and record when we last re-verified.
The same server, 5 different lenses. We reconcile these signals into our editorial score, which is why our number sometimes diverges from a directory-aggregate star count.
| Source | Their rating | Their star count | Their downloads | Last synced |
|---|---|---|---|---|
| AutomationSwitch This page | 4.5editorial | 37,989 | — | MAY 3, 2026 |
| PulseMCP | — unrated | unavailable | unavailable | APR 28, 2026 |
| MCPMarket | — unrated | unavailable | unavailable | APR 28, 2026 |
| Official MCP Registry | — unrated | unavailable | unavailable | APR 28, 2026 |
| Awesome MCP Servers | — unrated | unavailable | unavailable | APR 28, 2026 |
| Smithery | — unrated | unavailable | unavailable | APR 28, 2026 |
// Counts are directory-reported; we don't adjust them. Discrepancies usually come from different snapshot times or star-caching.
OTHER DEV TOOLS MCP SERVERS
GitHub
Official GitHub MCP server. 29,469 stars, MIT, 33 commits on main in the last 30 days. Repo, issue, PR, action, and code search tools backed by the GitHub REST + GraphQL APIs.
Playwright
Official Microsoft Playwright MCP server. Browser automation tools (navigate, click, fill, screenshot, evaluate JS, network capture) backed by Playwright. 31,942 stars, 35 commits/30d, Apache-2.0.
FastMCP (Python)
The fast, Pythonic way to build MCP servers and clients. Maintained by PrefectHQ. 25,133 GitHub stars and 92 commits on main in the last 30 days. The dominant Python MCP framework.
Repomix
Pack the entire repository into a single AI-friendly file. Optimised context for feeding codebases to Claude, ChatGPT, Gemini, and others. 24,515 GitHub stars and 100 commits on main in the last 30 days.
Serena
Powerful MCP toolkit for coding. Provides semantic retrieval and editing capabilities, positioned as the IDE for the agent. 23,971 GitHub stars and 100 commits on main in the last 30 days.
MCP Python SDK
The official Python SDK for Model Context Protocol servers and clients. Maintained by the Model Context Protocol project. 22,984 GitHub stars and 5 commits on main in the last 30 days.
DISCUSS YOUR
MCP REQUIREMENTS.
Evaluating a server, scoping an internal deployment, or working out whether MCP is the right fit at all. Start the conversation and we will point you at the right piece of the ecosystem.