Appearance
ScrapeGraphAI SOP
What It Is
ScrapeGraphAI is an AI-powered web scraping framework that converts webpages into structured data using large language models. It eliminates the need for CSS selectors or XPath expressions by letting you describe what to extract in plain language.
The library ships in two forms:
- Open-source library (
scrapegraphai) - self-hosted Python package that orchestrates LLMs (OpenAI, Ollama, Gemini, etc.) with Playwright to scrape and extract data locally. GitHub: ScrapeGraphAI/Scrapegraph-ai (27k+ stars). - Cloud API / managed platform - hosted service at scrapegraphai.com with SDKs (
scrapegraph-py,scrapegraph-js) and a REST API. Handles JavaScript rendering, proxy rotation, and anti-bot stealth automatically.
Category
Framework (AI-powered web scraping / LLM-directed data extraction)
Free vs Paid
Open-source library
Free and MIT-licensed. You supply your own LLM API key (OpenAI, Anthropic, Groq, etc.) or run a local model via Ollama. No usage fees from ScrapeGraphAI itself.
Cloud platform (scrapegraphai.com)
| Plan | Price | Credits |
|---|---|---|
| Free | $0 | 500 credits (one-time) |
| Starter | $20/month | 10,000 credits |
| Growth | $100/month | 100,000 credits |
| Pro | $500/month | 750,000 credits |
| Enterprise | Custom | Custom |
Credits never expire and stack on subscriptions. One-time top-up packs are also available ($5 for 1,000 credits up to $150 for 50,000 credits). Stealth mode costs +5 credits per request.
Official Links
- Website: https://scrapegraphai.com/
- Documentation: https://docs.scrapegraphai.com/
- GitHub (open-source): https://github.com/ScrapeGraphAI/Scrapegraph-ai
- MCP Server: https://github.com/ScrapeGraphAI/scrapegraph-mcp
Safety and ToS Notes
The open-source library README states: "ScrapeGraphAI is meant to be used for data exploration and research purposes only. We are not responsible for any misuse of the library."
Anonymous telemetry is collected by default in the open-source library. Disable it by setting the environment variable SCRAPEGRAPHAI_TELEMETRY_ENABLED=false.
The cloud platform is SOC 2 Type I compliant. Always review the target site's robots.txt and Terms of Service before scraping.