Evaluation record · mcp-server-puppeteer

MCP Puppeteer Server

v2025.4.6

Anthropic (Archived)

MCPbrowserautomationmcpmodel-context-protocol

Adequate

About This MCP

ARCHIVED: Former Anthropic reference MCP server for Puppeteer browser automation, archived 2025-05-29 to the servers-archived repository and no longer maintained. The archived repo explicitly provides no security guarantees. For browser automation, Microsoft's Playwright MCP server (microsoft/playwright-mcp) is the recommended successor and remains actively maintained as of July 2026.

Last Evaluated: July 9, 2026

Official Website

Trust Vector Analysis

Dimension Breakdown

🚀Performance & Reliability

browser automation accuracy

Automation task success testing

Evidence

Puppeteer Documentation — Reliable Chrome/Chromium automation with high success rates

highVerified: 2026-07-09

page load reliability

Page load success rate testing

Evidence

Puppeteer Network — Robust page loading with configurable wait strategies

mediumVerified: 2026-07-09

element interaction

Element interaction testing

Evidence

Puppeteer Selectors — Comprehensive selector support but depends on page structure

mediumVerified: 2026-07-09

screenshot quality

Screenshot quality assessment

Evidence

Puppeteer Screenshots — High-quality full-page and element screenshots

highVerified: 2026-07-09

error recovery

Error handling testing

Evidence

Error Handling — Good error reporting but complex failure scenarios require handling

mediumVerified: 2026-07-09

resource usage

Resource utilization monitoring

Evidence

Chrome Resource Usage — Significant memory and CPU usage for browser instances

mediumVerified: 2026-07-09

🛡️Security

browser sandboxing

Sandbox security review

Evidence

Chrome Sandbox — Chrome sandbox provides isolation but can be disabled

highVerified: 2026-07-09

arbitrary code execution risk

Code execution risk assessment

Evidence

Puppeteer Evaluate — AI can execute arbitrary JavaScript in browser context

highVerified: 2026-07-09

network access control

Network boundary testing

Evidence

Request Interception — Request interception available but AI can access any URL by default

mediumVerified: 2026-07-09

credential exposure risk

Credential security analysis

Evidence

Browser Automation — AI can interact with login forms and access authenticated sessions

highVerified: 2026-07-09

malicious site protection

Malicious content protection testing

Evidence

Chrome Safe Browsing — Chrome safe browsing provides some protection

mediumVerified: 2026-07-09

🔒Privacy & Compliance

web content exposure

Data exposure analysis

Evidence

Data Flow Analysis — Scraped web content and screenshots sent to LLM provider

highVerified: 2026-07-09

pii scraping risk

Privacy risk assessment

Evidence

Web Scraping Risks — No PII detection when scraping pages containing personal information

highVerified: 2026-07-09

cookie and session handling

Session security review

Evidence

Puppeteer Cookies — Can access and manipulate cookies, session data

mediumVerified: 2026-07-09

legal compliance

Legal compliance assessment

Evidence

Web Scraping Legality — Must respect robots.txt and terms of service; legal risks vary by jurisdiction

mediumVerified: 2026-07-09

tracking prevention

Privacy protection assessment

Evidence

Headless Mode — Headless mode provides some privacy but fingerprinting still possible

mediumVerified: 2026-07-09

👁️Trust & Transparency

documentation quality

Documentation completeness review

Evidence

Puppeteer Docs — Comprehensive Puppeteer documentation with examples

highVerified: 2026-07-09

action visibility

Action traceability assessment

Evidence

MCP Protocol Logging — Browser actions logged in MCP messages

mediumVerified: 2026-07-09

open source transparency

Source code review

Evidence

Puppeteer GitHub — Fully open source with Apache 2.0 license

highVerified: 2026-07-09

security guidance

Security documentation review

Evidence

Security Documentation — Limited security guidance for AI-controlled browser automation

mediumVerified: 2026-07-09

⚙️Operational Excellence

ease of setup

Setup complexity assessment

Evidence

Puppeteer Installation — Requires Chrome/Chromium installation and configuration

mediumVerified: 2026-07-09

resource efficiency

Resource utilization testing

Evidence

Browser Resource Usage — High memory and CPU usage, especially with multiple browser instances

highVerified: 2026-07-09

stability

Stability testing

Evidence

Puppeteer Reliability — Generally stable but can have issues with complex pages

MCP servers-archived repository — Server archived 2025-05-29; no bug fixes or compatibility updates will be released

mediumVerified: 2026-07-09

browser compatibility

Browser compatibility testing

Evidence

Puppeteer Browsers — Supports Chrome and Firefox through Chrome DevTools Protocol

highVerified: 2026-07-09

community support

Community activity analysis

Evidence

Puppeteer Community — Large community with 85k+ GitHub stars for the underlying Puppeteer library

MCP servers-archived repository — MCP server archived 2025-05-29; repository is read-only, issues and PRs are no longer accepted

Microsoft Playwright MCP releases — Recommended successor microsoft/playwright-mcp remains actively maintained with frequent releases (33k+ GitHub stars, releases through June 2026)

highVerified: 2026-07-09

Strengths

+Powerful browser automation for web scraping and testing
+Comprehensive API for page interaction and manipulation
+High-quality screenshot and PDF generation capabilities
+Large community and extensive ecosystem
+Open source with Apache 2.0 license
+Supports both Chrome and Firefox browsers

Limitations

!Significant security risk - AI can execute arbitrary JavaScript
!High resource consumption (memory and CPU)
!Can access credentials and authenticated sessions
!Scraped content sent to external LLM provider
!Legal and ethical concerns with automated web scraping
!No built-in controls for malicious site protection or PII filtering
!ARCHIVED 2025-05-29: unmaintained, no security guarantees; use Microsoft Playwright MCP instead

Metadata

license: Apache 2.0 (Puppeteer), varies for MCP implementation

supported platforms

0: Windows

1: macOS

2: Linux

programming languages

0: TypeScript

1: JavaScript

2: Python

mcp version: 1.0

browser support

0: Chrome

1: Chromium

2: Firefox

puppeteer version: 21.0+

resource requirements: High (500MB+ RAM per browser instance)

first release: 2024-11

maintained by: Unmaintained (archived 2025-05-29)

repository: https://github.com/modelcontextprotocol/servers-archived

Use Case Ratings

code generation

Useful for web scraping code generation and testing automation scripts

customer support

Can automate support workflows but limited direct applicability

content creation

Good for scraping content ideas, screenshots, and web research

data analysis

Excellent for web data extraction and competitive analysis

research assistant

Great for automated web research, data collection, and monitoring

legal compliance

Legal risks from automated scraping; terms of service violations

healthcare

Risk of scraping PHI from web portals; compliance concerns

financial analysis

Useful for market data collection but credential exposure risks

education

Good for educational web research and content aggregation

creative writing

Useful for research and inspiration gathering from web sources

Similar MCPs

Playwright MCP

Microsoft