// op_010 — case study In Dev
File Trail
// the problem
Important paper documents — IDs, insurance cards, receipts, contracts — end up as an unsorted pile of phone camera photos with no text search and no privacy gate. Finding one specific document later means scrolling through a camera roll, not searching a vault.
// architecture decisions
Capture & OCR
iOS Vision OCR extracts text from every captured document on-device, feeding a Claude-powered classification engine that auto-organizes files without the user manually filing anything.
Privacy
A Face ID-protected vault keeps sensitive documents gated behind biometric auth, separate from the general camera roll, with nested folders and smart renaming keeping the vault navigable as it grows.
Backend
Node.js/Express with PostgreSQL and Drizzle ORM, with an OpenAPI contract keeping the mobile client and server in lockstep as the schema evolves.
Monetization
RevenueCat gates the Pro tier — unlimited storage, AI categorization, and cloud backup — behind a subscription, keeping the free tier genuinely usable for light document storage.
// key outcomes
Bulk organize tools and Claude-driven auto-classification mean the vault stays organized without the user manually filing every capture — OCR makes every stored document full-text searchable later, turning a photo pile into an actual archive.
// related builds