Does it require Chrome to be installed?

No. Puppeteer downloads its own bundled Chromium on first run (~150 MB). That keeps the version pinned and your system browser untouched.

Can it take screenshots?

Yes. The server exposes `screenshot` as a first-class tool: full-page, viewport, or a specific selector. The model gets the image back as a base64 payload it can describe or reason over.

Can the agent fill forms and click buttons?

Yes. Alongside navigate and screenshot, the server exposes `click`, `fill`, `select`, `evaluate` (run arbitrary JS), and `wait_for_selector`. That is enough to drive most web flows end-to-end.

Does it run headless or with a visible browser?

Headless by default, which is right for servers and CI. If you want to watch the agent drive, set `PUPPETEER_HEADLESS=false` in the `env` block of your MCP config.

Will it work inside Docker?

Yes, but you need a base image with the Puppeteer system libraries (or use `ghcr.io/puppeteer/puppeteer`). On Alpine you will need `chromium` from the community repo plus `PUPPETEER_EXECUTABLE_PATH` pointing at it.

When should I prefer the Fetch MCP instead?

Whenever the content is served as static HTML. Fetch is an order of magnitude faster and cheaper. Reserve Puppeteer for sites that depend on JS, sessions, or user interaction — pricing calculators, dashboards, auth-gated docs.

Puppeteer

Name: Top MCPs — curated MCP directory
Creator: Top MCPs
License: https://creativecommons.org/licenses/by/4.0/

Full browser automation: navigate, click, screenshot, and scrape.

Score 48(?)Community (Hisma — fork of archived Anthropic reference)MIT 1Verified May 31, 2026Top MCPs for Browser Automation →

Install View on GitHub

Quick answer

What it does

Controls a headless Chromium browser via Puppeteer, enabling navigation, DOM interaction, form filling, screenshot capture, and JavaScript execution.

Best for

Web scraping with JavaScript
Form automation
Screenshot capture
End-to-end testing

Not for

Simple page reading
High-speed bulk scraping

Setup recipe

Pick your client, then follow the three steps.

Install

claude_desktop_config.jsonjson

{
  "mcpServers": {
    "puppeteer": {
      "command": "npx",
      "args": [
        "-y",
        "@hisma/server-puppeteer"
      ]
    }
  }
}

Paste under mcpServers. Fully quit and reopen Claude after editing.

CLI or .mcp.jsonshell

claude mcp add puppeteer -- npx -y @hisma/server-puppeteer

Run from your repo. Commit .mcp.json to share with your team.

.cursor/mcp.jsonjson

{
  "mcpServers": {
    "puppeteer": {
      "command": "npx",
      "args": [
        "-y",
        "@hisma/server-puppeteer"
      ]
    }
  }
}

Global path: ~/.cursor/mcp.json. Reload window after editing.

.vscode/mcp.jsonjsonc

{
  "servers": {
    "puppeteer": {
      "command": "npx",
      "args": [
        "-y",
        "@hisma/server-puppeteer"
      ]
    }
  }
}

VS Code uses the "servers" key (not "mcpServers").

~/.codeium/windsurf/mcp_config.jsonjson

{
  "mcpServers": {
    "puppeteer": {
      "command": "npx",
      "args": [
        "-y",
        "@hisma/server-puppeteer"
      ]
    }
  }
}

Open via Cascade → hammer icon → Configure.

cline_mcp_settings.jsonjson

{
  "mcpServers": {
    "puppeteer": {
      "command": "npx",
      "args": [
        "-y",
        "@hisma/server-puppeteer"
      ]
    }
  }
}

Open via the Cline sidebar → MCP Servers → Edit.

~/.continue/config.jsonjson

{
  "experimental": {
    "modelContextProtocolServers": [
      {
        "transport": {
          "type": "stdio",
          "command": "npx",
          "args": [
            "-y",
            "@hisma/server-puppeteer"
          ]
        }
      }
    ]
  }
}

Continue uses modelContextProtocolServers with a transport block.

~/.codex/config.tomlshell

# ~/.codex/config.toml
[mcp_servers.puppeteer]
command = "npx"
args = [
  "-y",
  "@hisma/server-puppeteer",
]

Codex uses TOML. Each server is a [mcp_servers.<name>] subtable.

~/.config/zed/settings.jsonjsonc

{
  "context_servers": {
    "puppeteer": {
      "command": {
        "path": "npx",
        "args": [
          "-y",
          "@hisma/server-puppeteer"
        ]
      }
    }
  }
}

Zed calls them "context_servers". Settings live-reload on save.

ChatGPT → Apps directorynone

Puppeteer doesn't ship a hosted HTTPS endpoint today. ChatGPT supports remote MCP servers only — to use this server in ChatGPT you'll need to deploy it to a public HTTPS URL first (e.g. via Cloudflare Workers or Vercel) or wait for an official remote build.

2
Set required secrets
No credentials required — this MCP runs over stdio without authentication.
3
Try a minimum working prompt
Screenshot a page and extract its headlines
Open https://news.ycombinator.com, take a full-page screenshot named `hn.png`, then return the top 5 headlines with their score and number of comments.
Tested with: Claude Desktop, Cursor.

Tools & permissions

Tool	Description	Args	Side effects
`puppeteer_navigate`	Navigate the browser to a URL.	url: string	Network
`puppeteer_screenshot`	Take a screenshot of the current page or a selector.	name: string, selector?: string, width?: number, height?: number	Read
`puppeteer_click`	Click an element by CSS selector.	selector: string	Exec
`puppeteer_fill`	Fill an input element.	selector: string, value: string	Exec
`puppeteer_select`	Select an option from a dropdown.	selector: string, value: string	Exec
`puppeteer_hover`	Hover over an element.	selector: string	Exec
`puppeteer_evaluate`	Run a JavaScript expression in the page context.	script: string	Exec

Security & scope

Access scope: Exec
Sandbox: Launches a headless Chromium process under the user account running the MCP. The browser has the same network and filesystem access the user has.
Gotchas: Chromium can execute arbitrary JavaScript on pages it visits — treat it like a browsing agent running with your credentials.
Downloaded files land in the user home directory unless `downloadPath` is configured.
Heavy memory footprint — one Chromium process per MCP launch, not reused across tool calls by default.

Agent prompt pack

— copy into Claude, Cursor, or ChatGPT.

Paste into Claude, Cursor, or ChatGPT. Edit the [brackets] before sending.

Recommend the best MCP servers for [task: e.g. browser automation work] in [client: Claude].

Constraints:
- Prefer tools that are [official | open-source | read-only] — pick what matters for my use case.
- Exclude MCPs that require [e.g. a paid plan, OAuth-only flows, remote-only transport].
- Return at most 3 picks, ranked.

For each pick include:
1. One-sentence rationale.
2. The ready-to-paste install snippet for my client.
3. Any required secrets I need to create before installing.

Cross-check the top-mcps.com listing: https://top-mcps.com/top-mcps-for-browser-automation

Compare Puppeteer against a real alternative. Swap the second MCP in [brackets] if you want a different match.

Compare Puppeteer MCP vs [Fetch MCP] for the following job: [describe the job, e.g. "let an agent create GitHub issues on bug triage"].

Judge them on:
- Setup time and complexity (what a new user hits first).
- Auth model (none / API key / OAuth 2.1) and credential risk.
- Transport (stdio / Streamable HTTP / SSE) and where the server runs.
- Required secrets and the blast radius if they leak.
- Operational risk in an unattended agent loop.
- Which one is "good enough" for a weekend prototype vs. production.

End with one sentence: which should I pick for my scenario, which is: [my scenario].

References:
- https://top-mcps.com/mcp/puppeteer
- top-mcps.com listing for Fetch

Asks the agent to install and verify. Works inside Claude Code, Cursor Agent, Codex CLI.

Install the Puppeteer MCP server for my [client: Claude] at the default config path for that client.

Use the exact install snippet published at https://top-mcps.com/mcp/puppeteer (fetch https://top-mcps.com/mcp/puppeteer.json for the canonical server.json if you can read URLs).

Before finishing:
1. Create the required secrets (no secrets) and put them in the appropriate env block — do not hard-code them.
2. Restart or reload the client so it picks up the new server.
3. Verify the server is connected (green / running state) and at least one tool is listed.
4. If anything fails, read the client's MCP logs and report the exact error — do not silently retry.

Confirm when done and list the tools the server now exposes.

Frequently asked questions

What changed

— 2 updates tracked.

May 31, 2026
Refreshed install snippets and fact sheet; verified for 2026.
Nov 25, 2024
Initial directory listing.

More Browser Automation MCPs

Other tools in the same category worth evaluating.

Playwright

Official

Official Microsoft browser automation across Chromium, Firefox, and WebKit.

browserautomationplaywrighttesting

5 minMedium

Browserbase

Official

Hosted, isolated Chromium runtime for AI agents that need a fresh browser per task.

browsercloudbrowserbaseautomation

5 minMedium

Stagehand

Official

AI-native browser automation: act, observe, and extract in plain English.

browserstagehandbrowserbaseai-automation

5 minMedium

Compared with Puppeteer

Side-by-side breakdowns for the choices people most often weigh against this MCP.

PuppeteervsBrave Search PuppeteervsFetch PuppeteervsE2B PuppeteervsPlaywright

Exploring Top MCPs for Browser Automation? See all Browser Automation MCPs →

Quick answer

What it does

Best for

Not for

Setup recipe

Install

Set required secrets

Try a minimum working prompt

Tools & permissions

Security & scope

Agent prompt pack

Full details

Frequently asked questions

Does it require Chrome to be installed?

Can it take screenshots?

Can the agent fill forms and click buttons?

Does it run headless or with a visible browser?

Will it work inside Docker?

When should I prefer the Fetch MCP instead?

What changed

More Browser Automation MCPs

Playwright

Browserbase

Stagehand

Compared with Puppeteer