DocPilot — AI Document Extraction & Review

DocPilot — AI Document Extraction & Structured Export

The challenge Finance teams repeatedly retype invoices, contracts, receipts, and statements into spreadsheets. Manual entry is slow, error-prone, and difficult to audit.

The solution I designed and built an end-to-end workflow for PDF, image, and text uploads. Documents are parsed into a strict structured schema, stored with visible success or error states, reviewed and corrected in the web dashboard, checked for possible duplicates, then exported to Excel or CSV.

Engineering highlights • Human-in-the-loop review instead of trusting model output blindly • Deterministic demo mode for local testing and CI without an API key • Duplicate detection that flags risk without deleting records automatically • Docker-ready FastAPI architecture with SQLite by default and PostgreSQL support

Full source code and setup instructions: https://github.com/LuciferLiu/docpilot