Server-Side AI Receipt Intelligence (Deep-OCR & Itemization) #446
Description
Type: Feature | Difficulty: Level 3 | Tags: Backend, OCR, Image-Processing
Description: Currently, receipt scanning is basic. This task involves building a server-side Intelligence Engine that not only performs OCR but also itemizes the receipt (splitting one receipt into multiple expense entries automatically).
Core Requirements:
Backend Logic: Integrate a server-side OCR engine (like Tesseract.js or a Vision API wrapper).
Deep Parsing: Use Regex or a local NLP model to extract not just the total, but individual line items (e.g., "Bread: ₹40", "Milk: ₹60").
Auto-Split Logic: Logic to automatically create individual Expense entries from one single receipt upload.
Frontend: A "Scan Preview" modal that shows the extracted text and allows the user to "Confirm All Items" with one click
Files to be changed:
services/ocrService.js
services/parsingService.js
routes/receipts.js
models/Expense.js
, public/receipt-upload.js,
public/index.html
, package.json.