Core Features

Receipt Scanning

Point your camera at any paper receipt and Spend Sense extracts the total, merchant, date, and line items — all processed entirely on your device using Apple Vision.

Best Practice

No receipt image is ever saved to your Camera Roll, sent to a server, or stored outside your encrypted local database. On-device OCR is a core privacy commitment.

How it works

Spend Sense uses Apple Vision Framework — Apple's built-in machine learning text recognition engine — to perform optical character recognition (OCR) directly on the device's Neural Engine. No third-party OCR service is used.

How On-Device OCR Works

📷Step 1

Camera Capture

Live preview. No photo saved to Camera Roll.

🧠Step 2

Apple Vision

Apple Vision Framework extracts text on-chip.

🏷️Step 3

Parse & Tag

Merchant, total, date, and items extracted.

🔒Step 4

Save Locally

Transaction written to encrypted on-device store.

Your receipt image is never saved, uploaded, or transmitted. Processing is entirely on your device.

Scanning a receipt

Open the scanner

From any screen, tap the camera icon in the toolbar or press the receipt icon when logging a transaction manually.

Align and capture

Hold the receipt flat and well-lit. Tap the shutter button. The live viewfinder highlights detected text regions in real time so you know before you tap that the text is readable.

Review extracted data

Spend Sense presents a pre-filled transaction form with the extracted total, merchant name, and date. Verify and correct any field before saving.

Attach items (optional)

If line items were detected, you can review them as sub-items attached to the transaction — useful for splitting bills or tagging individual products.

Tip

For best results: lay the receipt flat on a dark surface, ensure the full receipt is in frame, and scan in good lighting. Thermal fax-paper receipts fade quickly — scan them soon after printing.

What gets extracted

Field	Source on Receipt	Accuracy
Total amount	Bottom total / Grand total line	High — usually largest number
Date	Header or footer timestamp	High — multiple format patterns
Merchant name	Header, logo text, or store name	Medium — depends on font/format
Individual line items	Itemized list in body	Medium — works best on structured receipts
Tax & tip	Subtotal section	High when clearly labeled
Payment method	Footer (VISA, CASH, etc.)	Low — highly variable formatting

Solving common problems

Problem	Cause	Fix
Scanner misses text	Low contrast or glare	Move to diffuse light; avoid direct sun
Wrong total extracted	Multiple totals on receipt (tax, subtotal)	Manually correct in review screen before saving
Merchant name blank	Logo-only header (image, not text)	Type the merchant name after scanning
Date shows wrong year	Receipt only shows MM/DD	Spend Sense defaults to current year — correct if needed
Receipt too crumpled	Distorted text geometry	Flatten against a flat surface before scanning

Privacy architecture

On-device processing only

All OCR runs on Apple Vision — a framework built into iOS, not a cloud API.

No image retention

The captured frame is discarded immediately after text extraction. Nothing is written to disk.

Encrypted local storage

Extracted transaction data is stored in an AES-256 encrypted SQLite database.

No telemetry

Spend Sense does not send analytics about your scans, failure rates, or merchant names.

Info

Spend Sense requires camera permission to scan receipts. This permission can be revoked at any time in iOS Settings → Privacy & Security → Camera without affecting any other feature.

Was this page helpful?

Yes, thanks!Needs improvement

Quickstart

Budgeting