This is a full-stack document processing app built for the CDIS × Endeavor AI Hackathon. It allows users to upload real-world documents (like purchase orders), extract structured data using Unstruct AI, and manually verify or edit the output.
Note: Since it is a private repo, I have directly used the keys instead of saving it in a .env file. This will save the time for the judges.
- Upload PDFs, PNGs, DOCX, and other common document formats
- View and trigger processing from the Progress page
- Verify, edit, or reject extracted fields on the Verification page
- Re-send a document for reprocessing if needed
- One-time landing animation for a polished intro
- Fully local — all data is stored and handled on your machine
| Layer | Technology |
|---|---|
| Frontend | React, React-Bootstrap, Axios, React Router |
| Backend | Node.js, Express, Multer, Axios |
| AI Service | [Unstract AI] |
| Storage | Local filesystem (uploads/, structuredOutputs.json) |
cd backend
npm install
# Create uploads folder structure if it already doesn't exist
mkdir -p uploads/new uploads/processed
echo "[]" > uploads/structuredOutputs.json
echo "[]" > uploads/uploadedFiles.json
npm index.js
cd ../frontend
npm install
npm run dev