What it does
An open-source engine that reads documents such as invoices, contracts or receipts and extracts the fields you define in a schema. The results are stored as typed records, so you can filter and add them up with exact results. Questions are asked in plain language, and the project includes an MCP server and Python and TypeScript SDKs. It can be self-hosted.
Use cases
- 01Totalling amounts across a folder of supplier invoices
- 02Listing key dates and parties from signed contracts
- 03Filtering receipts by client or period for review