Importing a payslip (with the privacy review step)
Payslip import turns a photo of your payslip into a monthly salary record — without the image or its text leaving your device.
The flow
- Salary → import payslip, then pick one or more images from your photo library. (PDF is not supported — images only.)
- Location and other metadata (EXIF) is stripped from the images immediately.
- Privacy review. Before any text extraction, you see each page with a crop and redact editor. Varuno’s sensitive-data detection highlights content that looks like an NRIC or similar identifier so you can black it out. This step is mandatory, not optional.
- On-device extraction. Text recognition runs entirely on the phone and Varuno extracts fields such as gross pay, net pay and CPF amounts.
- Extraction review. Every extracted field is shown and editable. Nothing is saved until you explicitly confirm.
- On save, the monthly salary record is written to your encrypted database. The working images are deleted — however the import ends, the session workspace is cleaned up, and abandoned imports are swept after 12 hours.
Keeping a copy
At the review step you can tick “Keep a copy of this document on this device”. This keeps the processed pages — after your crops and redactions, not the originals — in the app’s local storage. You can view or delete kept documents under Settings → Data → Local Storage.
If OCR finds nothing
Low-quality photos and non-Latin scripts are the usual causes (recognition currently handles Latin script). Retake the photo flat and well lit, or use the manual entry fallback, which is always available. If OCR never works at all on your device, report it with your device model — and never send the payslip itself unredacted.
A note on honesty
This pipeline is fully implemented and heavily tested in simulation, but on-device verification is still in progress as part of the current TestFlight QA cycle. If something misbehaves, please report it — early reports are exactly what TestFlight is for.