🔍 zenrows-skill

Web scraping skill for OpenClaw agents — powered by ZenRows.

When web_fetch fails due to anti-bot protection, JS-rendered content, or geo-restrictions, ZenRows handles it:

• Anti-bot bypass — Cloudflare, DataDome, PerimeterX, etc.
• JavaScript rendering — SPAs, dynamic content, lazy-loaded pages
• CAPTCHA solving — Automatic CAPTCHA resolution
• Proxy rotation — Residential proxies from 190+ countries
• Output formats — HTML, Markdown, Plaintext, PDF, Screenshots

---

Quick Start (OpenClaw)

1. Install:

   clawhub install Penguin-Life/zenrows-skill

2. Add your API key to openclaw.json:

   {
     "env": {
       "vars": {
         "ZENROWS_API_KEY": "your-api-key-here"
       }
     }
   }

   Get a key at zenrows.com (free tier: 1,000 credits).

3. Restart OpenClaw:

   openclaw gateway restart

4. Ask your agent:

   "Fetch https://protected-site.com as markdown"

---

Standalone CLI

# Basic fetch
./zenrows/scripts/zenrows.sh "https://example.com"

# Markdown output (ideal for LLM context)
./zenrows/scripts/zenrows.sh "https://example.com" --markdown

# Auto mode (ZenRows picks optimal anti-bot strategy)
./zenrows/scripts/zenrows.sh "https://protected-site.com" --auto

# JS rendering for SPAs
./zenrows/scripts/zenrows.sh "https://spa-app.com" --js --wait-for ".content"

# Extract structured data with CSS selectors
./zenrows/scripts/zenrows.sh "https://news.site.com" \
  --css '{"title":"h1","body":".article-body"}'

# Specific country proxy
./zenrows/scripts/zenrows.sh "https://geo-restricted.com" --premium --country us

Example Output

$ ./zenrows.sh "https://example.com" --markdown

# Example Domain

This domain is for use in illustrative examples in documents.
You may use this domain in literature without prior coordination...

---

When to Use ZenRows vs web_fetch

Need web content?
├── Try web_fetch first (free)
│   ├── ✅ Success → done
│   └── ❌ Blocked / empty / broken
│       ├── Static site, basic protection → zenrows (1 credit)
│       ├── JS-heavy SPA → zenrows --js (5 credits)
│       ├── Strong anti-bot → zenrows --auto (variable)
│       └── Structured extraction → zenrows --css / --autoparse
└── Need markdown/text for LLM? → zenrows --markdown / --text

Credit Costs

Config	Credits
Basic (HTML only)	1
JS rendering (--js)	5
Premium proxies (--premium)	5
JS + Premium	10
CAPTCHA solving (--captcha)	+25
Auto mode (--auto)	Variable

CLI Options

Option	Description
--auto	Adaptive Stealth Mode
--js	Enable JS rendering
--premium	Residential proxies
--country <CC>	Proxy country (implies --premium)
--markdown	Markdown output
--text	Plaintext output
--pdf	PDF output
--autoparse	Auto-extract structured data
--css '<json>'	CSS selector extraction
--wait <ms>	Wait fixed ms (requires --js)
--wait-for <sel>	Wait for CSS selector (requires --js)
--captcha	Auto-solve CAPTCHAs
--screenshot	Base64 screenshot
--json	JSON response with headers
-o <file>	Write to file
-v	Verbose mode

See zenrows/SKILL.md for the full API parameter reference.

---

Project Structure

zenrows-skill/
├── skill.json              # Package manifest
├── README.md               # This file
├── zenrows/
│   ├── SKILL.md            # Full agent instructions + API reference
│   └── scripts/
│       └── zenrows.sh      # CLI wrapper
└── .github/workflows/
    └── ci.yml              # shellcheck + smoke tests

License

MIT

---

Built with 🐧 by Penguin-Life