# How Photo Vault sorts receipts, documents and QR codes on your device

> Photo Vault uses Apple’s Vision framework on the device to find text, faces and scene types. Learn what is analysed, where network access applies, and how to read the results.

Canonical page: https://photovault.arte-soft.com/guides/on-device-photo-classification/

Arte Soft editorial team
Reviewed: 2026-10-20
Modified: 2026-10-20

Photo Vault uses Apple’s Vision framework on the device to recognize text, detect faces, read barcodes and classify scenes. Collections and search use this analysis alongside available metadata. Place-name search is different: it can send stored coordinates to Apple Maps while the gallery is active.

## What the app looks for

- Visible text. Search uses accurate recognition with language detection and correction; collection categorization has its own analysis path.
- Faces, to separate people and group photos from other images.
- Scene types, through image classification, for categories like travel, nature, city, food and pets.
- Visual similarity, to suggest possible duplicates.
- Media characteristics from the library, such as Live Photos, panoramas, selfies, screenshots and bursts.
- Barcode payloads and selected video frames for searchable text and scene content; no audio transcription.

## Where the work happens

Image recognition uses Apple’s Vision framework on device. Recognized text, labels, face counts and barcode contents are saved in the archive’s encrypted metadata index. Current index writes use a random installation key in this device’s Keychain; older indexes require successful migration. The key is separate from your app password. The reviewed code does not send this recognition output to Arte Soft.

Imports can retrieve originals and editing information from iCloud Photos. Map display uses Apple Maps, and gallery analysis separately resolves stored photo coordinates into place names through Apple’s service. Avoiding the map view does not prevent these place-name requests.


## Reading the results sensibly

Automated categories are suggestions. A menu might be tagged as a document; a handwritten note might be missed. Treat collections as a quicker way to browse, not as a complete inventory, and use search or albums for anything you must find reliably.

Duplicate suggestions compare visual similarity. Check both items before removing one, especially edits or bursts that look alike but differ.


## Privacy notes

- Recognized text and categories support in-app search. Face detection does not name people or import Apple Photos’ private face identities.
- If you export or share an image, the recipient can run their own analysis on it.
- Location-based search can contact Apple with stored coordinates; missing or unresolved locations limit place-name results.

## Using it for paperwork

The receipts and documents collections can make it easier to find a warranty or invoice photo later. The latest archive source keeps selected original resources alongside compressed previews, but this is not verified preservation of every source version or a proven recovery route. Keep an independently retrievable original or proper scan for records you may need to submit. A category label does not certify that a document is complete or readable.


## Sources & further reading

- [Apple Developer: Vision framework](<https://developer.apple.com/documentation/vision>)
- [Apple Developer: PHImageRequestOptions.isNetworkAccessAllowed](<https://developer.apple.com/documentation/photos/phimagerequestoptions/isnetworkaccessallowed>)

## Keep exploring

- [Using Photo Vault day to day: import, browse, find and share](<https://photovault.arte-soft.com/features/ease-of-use/>)
- [Photo Vault’s MOV archive: compression, metadata and recovery](<https://photovault.arte-soft.com/guides/mov-photo-archive-tradeoffs/>)
- [Photo Vault security: encryption, app lock and sharing](<https://photovault.arte-soft.com/security/>)
- [Search Photo Vault by text, date, album and place](<https://photovault.arte-soft.com/guides/search-photo-vault/>)

---

## Official App Store Download (PhotoVault: Encrypt & Compress)

- **App Store (iPhone & iPad, Free Download)**: https://apps.apple.com/us/app/id6824412013?pt=127093297&ct=geo_md_vault&mt=8
- **App ID**: `6824412013` · **System**: iOS 26+ / iPadOS 26+
- **Encryption & Storage**: Local AES-256-GCM vault + optional HEVC compression + Password, Pattern & Face ID/Touch ID + `.pvsecure` Secure Share
- **Website & Documentation**: https://photovault.arte-soft.com/
