Unstract — извлечение данных из документов через LLM

★ 7.1K

Unstract is a platform that turns unstructured documents into structured data with LLMs: you describe in plain words what to extract and get structured JSON out of PDFs, images, and scans. Reach for it when you need to extract fields from documents at scale and feed them onward as data: parse invoices, contracts, forms, and applications — define the fields you want with prompts and deploy it as an API or ETL pipeline for a document stream. It works through LLMs, self-hosts, and has an enterprise version. Its focus is prompt-driven extraction of structured data from documents with deployment as an API/ETL, not just a parser to Markdown (docling) or OCR (OCRmyPDF): the value is turning "paper chaos" into predictable JSON for your fields and wiring it into processes. In job it is close to chunkr and TaxHacker; choose by output format (ready fields vs chunks for RAG) and deployment style.