unstract
LLM-Driven Extraction of Unstructured Data — Built for API Deployments & ETL Pipeline Workflows
What is unstract?
LLM-Driven Extraction of Unstructured Data — Built for API Deployments & ETL Pipeline Workflows
Overview of unstract
What it is
Unstract is an open-source tool that uses LLMs to extract structured JSON from unstructured documents such as PDFs, images, and scans. Users define extraction schemas using natural language prompts and deploy the extraction as a REST API or an ETL pipeline. It is designed for teams in finance, insurance, healthcare, KYC/compliance, and other industries.
Key features
- Prompt Studio — Define document extraction schemas with natural language.
- API Deployment — Send a document over REST API and receive JSON back.
- ETL Pipeline — Pull documents from a folder, process them, and load to a data warehouse.
- MCP Server — Connect to AI agents (e.g., Claude) via Model Context Protocol.
- n8n Node — Integrate into existing automation workflows.
Use cases
Unstract is used for document data extraction in finance and banking, insurance, healthcare, income tax, and KYC/compliance workflows. It handles various document formats including PDF, DOCX, images, spreadsheets, and presentations.
Getting started
git clone https://github.com/Zipstack/unstract.git
cd unstract
./run-platform.sh
License and pricing
Unstract is released under the AGPL-3.0 License and is open source.
Based on the project's public README and repository metadata.