🛡️ Playtrafi

The undetected stealth web scraper & clean Markdown extractor for LLMs and AI pipelines.

⚡ What is Playtrafi?

Playtrafi (formerly Patchtroy) is a lightweight, high-performance web crawling and content extraction engine designed specifically for LLMs, RAG pipelines, and AI agents.

It bridges the gap between low-level drivers and high-level scrapers by pairing Patchright (the C++ patched Chromium engine that eliminates automated telemetry leakages) with Trafilatura (the gold standard for algorithmic boilerplate removal and pristine Markdown conversion).

🎯 Why Playtrafi vs Crawl4AI & Firecrawl?

📦 Installation

pip install playtrafi

# Install stealth Chromium driver
patchright install chromium

Note on Backwards Compatibility: patchtroy remains available as an import and CLI alias during the migration deprecation window.

🚀 Quickstart

Python (Async API)

import asyncio
from playtrafi import AsyncPlaytrafi

async def main():
    async with AsyncPlaytrafi() as client:
        result = await client.scrape("https://news.ycombinator.com")
        print("Title:", result.title)
        print("Markdown:\n", result.markdown[:300])

async def_run():
    asyncio.run(main())

Python (Sync 1-Liner)

from playtrafi import Playtrafi

result = Playtrafi.crawl("https://en.wikipedia.org/wiki/Web_scraping")
print(result.markdown[:300])

CLI

# Scrape clean Markdown directly to stdout
playtrafi https://news.ycombinator.com

# Save to Markdown, CSV, or JSON (auto-detected from file extension)
playtrafi https://example.com -o output.md
playtrafi https://example.com -o output.csv
playtrafi https://example.com -f json -o output.json

🧩 Core Architecture Highlights

• 🛡️ Native Stealth Engine: Runs on Patchright's patched C++ Chromium binary, eliminating internal debugging flags and CDP telemetry before page scripts execute.
• 🧹 Algorithmic Content Cleaning: Trafilatura strips 96%+ of navigation menus, advertisements, and tracking noise while preserving tables, code snippets, and Markdown formatting.
• ⚡ Browser Context Pool: Reuses a single background browser process across requests, leasing isolated browser contexts for >5x higher throughput and lower RAM usage.
• 🧠 Structure-Aware LLM Chunking: Recursively splits Markdown documents along heading hierarchy (#, ##, ###) with token estimation and metadata tracking.
• 🐳 Standalone REST Microservice: Single lightweight Docker container exposing /scrape, /scrape/batch, and /health on port 4013 for any programming language.

📖 Documentation

Full documentation, API references, architecture guides, and benchmarks are available at the Playtrafi Documentation Portal:

• Getting Started Guide
• Stealth Architecture Deep Dive
• Context Pooling & Concurrency
• Media Capture (Screenshots & PDFs)
• Proxy Rotation & Fault Quarantine
• LLM Chunking & Tokenizer
• REST API Microservice & Docker
• Benchmark Results vs Crawl4AI & Firecrawl
• Rebranding Guide

📄 License

Playtrafi is open-source software created by @marcuz-apl and maintained by Alfazen Inc., released under the Apache 2.0 License.