Dainty Trading
Home / Projects / ReceiptSnap AI

ReceiptSnap AI

Native iOS and Android receipt scanner with on-device capture, cloud OCR, and a tidy expense log. Designed for solo operators who want speed at the till and structured data afterwards, without an SaaS subscription.

Category  AI · Mobile Stack  Swift/SwiftUI · Kotlin/Jetpack Compose · Firebase LLM  Google Cloud Vision Status  In development

The problem

Receipt-tracking apps either bury the user in features they don’t want (multi-currency travel reports, team approval flows, subscription tiers) or are so basic they don’t actually save the data anywhere structured. There’s a gap for solo operators — tradies, freelancers, sole traders — who want a fast camera button and a clean export at year-end.

What we’re building

ReceiptSnap AI is a native app on both platforms (Swift/SwiftUI on iOS, Kotlin + Jetpack Compose on Android). The capture flow is optimised for one-handed use at the till. Google Cloud Vision runs OCR; results sync to Firebase; an export bundles the year as CSV plus original images for your accountant.

The AI angle

OCR alone isn’t enough — the value is in the post-OCR cleanup. We extract structured fields (vendor, date, total, GST, line items) and confidence-score each one. Low-confidence fields surface for one-tap correction. Over time the model learns the user’s common vendors and pre-fills with high accuracy.

How it’ll be used

  • Tradies and contractors capturing job-related expenses on the go.
  • Freelancers and sole traders who do their own books.
  • Travel-heavy professionals who want every receipt logged before the day ends.

Where we are

iOS build is in TestFlight. Android build is following close behind. Public release is gated on the year-end export hitting the format conventions of common accounting software.