Shipping a mobile utility that offloads the entire document ingestion and OCR inference pipeline to Apple's native Vision framework on-device is an exceptional architectural choice. Most modern scanning applications treat basic document capture as a hook to force user registration, track analytics, and demand recurring cloud subscriptions. Clearcan eliminates this operational friction completely by executing local camera capture, automated canvas edge-detection, and text recognition entirely within the device's sandbox. The privacy advantages are absolute—it completely clears enterprise data compliance and privacy liabilities, as sensitive financial or personal papers never traverse an external server. Furthermore, performance is incredibly snappy because running hardware-accelerated machine learning models locally on the Neural Engine bypasses network round-trip latency entirely.
While leveraging Apple's local Vision framework keeps data strictly isolated, the multilingual text recognition capabilities are currently restricted to a limited lexicon (English, Russian, and Romanian). Expanding the on-device NLP language modules to handle broader character scripts or complex diacritics without inflating the application’s binary footprint is a key engineering hurdle that needs addressing. Additionally, running high-resolution, multi-page batch scanning alongside real-time OCR text serialization sequentially on older iOS hardware will inevitably push local memory constraints, risking thread termination if background resource usage spikes. The search repository layer would also benefit from more granular data management features—such as user-defined metadata tags or structured folder hierarchies—to prevent a massive local document database from becoming unmanageable over time.
I’ve previously balanced bloated, subscription-based platforms like CamScanner or Adobe Scan against the native document scanner tucked inside the default Apple Notes app. Commercial utilities have become heavily compromised by aggressive paywalls, forced cloud sync pipelines, and tracking analytics that create massive data liabilities. On the other hand, the scanning tool inside Apple Notes handles basic capture well but lacks a dedicated file-management system, advanced multilingual text extraction workflows, and streamlined standalone PDF export options. I opted for Clearcan because it treats document digitization as a pure, lightweight utility pipeline—respecting local privacy boundaries and system resources without locking essential tools behind an artificial payroll gridlock.

@basogulgon30432 Hi! Long documents are fine — recognition runs per page at save time, around half a second each, so a 20-page contract is no different from a single receipt. Handwriting is weak, as I told Rabia above — printed text is where it shines.
On languages: English, Russian and Romanian are what I actively tested (including Romanian diacritics). Apple's on-device models cover many more, so I'm planning to test and enable additional ones — if you tell me which language you need, I'll prioritise it.
@rabia597613 Hi Rabia! It uses Apple's Vision framework on-device, so printed text — including crumpled receipts, shadows, odd angles — is handled well. Handwriting is the honest weak spot: block letters sometimes work, cursive mostly doesn't.
the "everything stays on your phone" pitch is great until you factor in what happens when the phone dies or gets replaced. no cloud is a real privacy win but it also means no cloud backup by default, is there any local export/backup path (iCloud folder, encrypted file, anything) so a lost phone doesn't mean losing every scanned document, or is that an intentional tradeoff you're fine with?
@galdayan Fair point, and an honest gap. Right now it's manual: any document exports to PDF via the share sheet, and from there you can Save to Files — which can live in iCloud Drive if you want. So you can back things up, but only if you remember to.
A proper local backup is on my list — one file you own, saved wherever you like. Still no cloud, still no account. I'm building this step by step.
Thanks for the push!
@ion_hortolomei 'saved wherever you like, no cloud, no account' is the right shape for the backup feature too, keeps it consistent with everything else you've built. building it step by step instead of bolting on a cloud sync to solve it fast is the harder but more honest path.
@galdayan Yeah, that's the plan. Backup will just be a file you own — no server, no account.
@ion_hortolomei that's exactly the right shape - a file you own beats an account you have to trust. good luck shipping it, will keep an eye out for the update.
Love the no-subscription stance, that's a breath of fresh air for scanner apps. One thing that would make this a daily driver for me: support for splitting a multi-page scan into separate documents automatically, or at least let me reorder and delete pages right inside the app before exporting the PDF.
@irmak2070596 Hi Irmak! Deleting pages actually already works — tap the trash icon on any page, both while scanning and inside a saved document.
Reordering is genuinely missing and it's next on my list. Splitting one session into separate documents is a good idea too — noted. Thanks, this is the useful kind of feedback.
@sabriboruc9hqs Hi! Actually Cyrillic was the whole reason I went with Apple Vision instead of ML Kit — ML Kit still doesn't do Cyrillic at all. Good to hear it works on your real documents)





Thank you for app review) Here i built the app I was looking for myself: point, scan, get a PDF. That's it.
Two notes about improvements: expanding languages actually costs nothing in binary size — Vision's models ship with iOS, not with the app (that's exactly why I picked it over ML Kit, which does bundle models and still has no Cyrillic). So more languages are mostly a matter of testing, and that's on my list.
Your point about memory on multi-page batches with older hardware is the sharp one — I'll test that properly.
Thanks again for the feedback!