Invisible_playwright_mcp
Invisible_playwright_mcp is an open-source Model Context Protocol (MCP) server that exposes an anti-detect, "stealth" browser to AI agents and automation scripts. It pairs the Playwright automation API with browser fingerprint patching so that automated sessions look like ordinary human browsing, reducing the frequency of bot-detection challenges and CAPTCHA interstitials encountered during scripted navigation. The project is positioned as an "agentic stealth browser": an MCP-compatible tool layer that LLM clients and Python agents can call to drive a real browser. Typical operations include navigating to URLs, reading rendered page content, clicking and filling forms, extracting structured data, and following multi-step workflows, all through the standardized MCP tool interface rather than bespoke glue code. Because it is delivered as an MCP server, it is designed to plug into MCP-capable clients (such as AI coding assistants and agent frameworks) with minimal configuration. The underlying automation is Python-based, so it also fits naturally into existing Python data, scraping, and research pipelines. The repository markets the tool around undetected browsing, browser automation, computer-use style agents, web scraping, and lead generation, with the stated benefit of encountering fewer CAPTCHAs. License, star count, and release cadence are not asserted here because they were not verifiable from the supplied metadata.
- Category
- Productivity
- Pricing
- Open Source
Problem solved
Standard headless browser automation is easily fingerprinted, so scrapers and AI browsing agents frequently hit bot-detection walls, CAPTCHAs, and blocked sessions. This project wraps Playwright in stealth/anti-detect behavior and exposes it through MCP, letting agents drive a browser that more closely resembles normal human traffic and reducing the engineering effort needed to keep automated browsing sessions unblocked.
Use cases
- Giving an LLM or MCP-compatible agent the ability to browse and read live web pages
- Scraping JavaScript-heavy sites that render content client-side
- Automating multi-step form filling and navigation without triggering bot checks
- Lead generation and prospect research across public web directories
- Building computer-use style agents that click, type, and extract on real pages
- Running undetected regression or smoke tests against a production website
- Collecting web data for research datasets from sites with anti-bot defenses
Audience
- Developers
- ML Engineers
- Data Scientists
- Researchers
- Startups
Pros
- Stealth/anti-detect layer reduces bot detection compared with vanilla Playwright
- MCP interface makes it easy to attach to MCP-capable AI agents and clients
- Python-based, so it drops into existing scraping and data pipelines
- Supports real rendered-page interaction, including clicks, typing, and extraction
- Free and open source, with no vendor subscription
Cons
- Anti-detect tooling sits in a gray area with many sites' terms of service
- Setup typically requires local browser binaries and MCP client configuration
- License, maintenance activity, and release cadence were not verifiable from the provided metadata
- Stealth techniques can break when target sites change their detection methods
- No guaranteed CAPTCHA solving despite the "no captchas" marketing phrasing
Official website: https://github.com/feder-cr/invisible_playwright_mcp
Canonical LambdaBase page: https://www.lambdabase.com/tools/invisible-playwright-mcp