Project
Roulette

Expense Ledger

Photograph a receipt, check what it read, and file it.

AdvancedWeb appProductivity16–32hAI-buildable 3/5no setupworth shipping

Capture receipts, run OCR in the browser, and extract merchant, date and total into a reviewable ledger. OCR is never fully reliable, so the design has to make correcting a misread faster than typing it fresh — that review step is the whole product.

UI
Logic
Useful

Hosting · Free forever

Deploys free on GitHub Pages, Cloudflare Pages or Vercel Hobby.

Deploy it and share a link.

Core features · 7

  • Capture or upload a receipt image
  • OCR extracts text in the browser
  • Merchant, date and total are parsed from the text
  • A review step shows extracted values beside the image for correction
  • Low-confidence fields are flagged for attention
  • Expenses categorise and total by period
  • Images and data stay local and export together

The interesting part

  • OCR accuracy on crumpled thermal paper is genuinely poor
  • Parsing dates and totals from wildly varying receipt layouts
  • Making correction faster than re-entry, or the feature is pointless

Once it works

  • Add line-item extraction
  • Add currency detection for foreign receipts
  • Add duplicate receipt detection

Suggested stack

reacttypescripttailwindtesseract.jslocalstorage

Track it