What is LLM Scraper?
LLM Scraper is an open-source TypeScript/Node.js library that extracts structured data from webpages using LLMs. It ships as the `llm-scraper` npm package under the MIT license. It drives Playwright for browser automation and JavaScript rendering, then sends page content (HTML, markdown, plain text, a screenshot, or a custom format) to a connected provider (OpenAI, Anthropic, Google Gemini, Groq, or a local Ollama model) to fill a schema the developer defines with Zod or JSON Schema. It also supports streaming output and schema-to-code generation. There's no dashboard, account system, or hosted pricing: it runs entirely inside the developer's own codebase.
Our verdict
This suits developers who already use Playwright and an LLM API key and want direct control over extraction schemas in a TypeScript codebase, rather than teams that want a hosted API with a dashboard and support. Around 6,900 GitHub stars and roughly 108 commits from a single maintainer suggest real developer adoption, but the project has a thinner commit history and less institutional backing than commercially funded extraction APIs.
Categories:
AI extraction tools sit between a raw page and your database: you describe the fields you want, and a model returns structured JSON or LLM-ready markdown without hand-written selectors. They hold up well when page layouts change often, at the cost of per-request model spend and the occasional wrong field.
