ScrapeGraphAI
A Python library that uses an LLM and graph-based pipelines to scrape structured information from websites and local documents such as XML, HTML, JSON and Markdown, from a plain-language description of what to extract.
The library offers several ready-made scraping pipelines: SmartScraperGraph extracts information from a single page or file given a prompt and a source; SearchGraph extracts information from the top results of a search-engine query; SpeechGraph extracts information from a page and turns it into an audio file; ScriptCreatorGraph generates a Python script that performs the extraction instead of running it directly; and multi-page variants of each pipeline extract from several sources at once, optionally running LLM calls in parallel.
It works with LLM providers including OpenAI, Groq, Azure, Gemini and MiniMax, or with local models through Ollama, and needs Playwright installed separately for fetching website content. It integrates with frameworks such as LangChain, LlamaIndex, Crew.ai and Agno, low-code platforms such as Pipedream, Zapier and n8n, and exposes an MCP server alongside Python and Node SDKs.
It is licensed under the MIT license.