unstract
Developed by Zipstack
Unstract is an open-source platform by Zipstack that leverages LLMs to extract structured JSON data from unstructured documents like PDFs, images, and scans. It allows users to define extraction schemas using natural language prompts in its 'Prompt Studio' and deploy them as REST APIs or ETL pipelines. It features Model Context Protocol (MCP) server support, n8n integration, and broad compatibility with major LLM providers, vector databases, and data warehouses.
- Natural language prompt-driven document extraction
- Out-of-the-box ETL pipelines and REST API deployment
- Model Context Protocol (MCP) server support for AI agents
- Highly compatible with major LLM providers and vector DBs
web