ScrapeGraphAI

ScrapeGraphAI extracts data from websites and documents based on a plain-language request. Developers can use its self-hosted Python library or call a separate managed cloud API.

• Prompt-based extraction: Describe the information you need from a page instead of writing page-specific selectors.
• Local documents: Extract information from XML, HTML, JSON and Markdown files as well as website content.
• Model choice: The library supports local LLMs through Ollama and cloud providers including OpenAI, Groq, Gemini and Azure.
• Application integrations: Connect extraction to LangChain, LlamaIndex and Crew.ai, or workflow tools such as n8n, Dify and Zapier; an MCP server is also available.

The open-source library uses the MIT license and runs on your infrastructure. Ollama keeps model inference local, while cloud model choices send requests to the selected provider. Scraping a website still requires access to that site. The package collects anonymous usage metrics, which you can disable.

The hosted API runs in ScrapeGraphAI's cloud and manages the LLM, browser rendering, proxies and anti-bot handling. Those managed capabilities belong to the cloud service; the self-hosted library uses your own browser and scraping setup. Python and JavaScript/TypeScript SDKs connect applications to the hosted API.