On-device processing handles data locally on the device. SnapAction stores resource cards locally with SwiftData, though AI analysis requires backend processing.
On-device processing means data is handled locally on the user’s device rather than sent to remote servers.
What SnapAction does on-device
- Stores resource cards, tags, metadata, and scan records locally with SwiftData
- Queries and searches the resource library
- Displays Rewind daily digests from the local index
- Runs the SwiftUI interface
What requires backend processing
- AI screenshot analysis through OpenRouter
- Canonical URL recovery through Serper search
- Subscription entitlement verification through Clerk and Convex
Why the split matters
Local storage keeps your resource library fast, available offline, and under your control. Backend processing is required for the AI analysis that turns screenshots into structured cards. SnapAction is transparent about this split and does not claim fully offline AI analysis.
How to evaluate privacy claims
On-device processing is a specific architecture claim. A product should be clear about which steps happen locally and which steps require a server, model provider, search provider, identity provider, or subscription system. Saying “local” is not enough if important parts of the workflow still leave the device.
For screenshot tools, the distinction matters because screenshots may contain private messages, receipts, travel details, work documents, or account information. Users should know whether the app stores the library locally, whether AI analysis is remote, and which screenshots they should avoid scanning.
Related concepts
- SwiftData stores structured local data in SnapAction.
- Vision AI can be local or remote depending on the implementation.
- What happens to my screenshots? explains the SnapAction workflow.
Sources
- Apple Developer: Protecting user privacy
- Apple Machine Learning
- Last checked: 2026-05-09
FAQ
What is On-device processing?
On-device processing handles data locally on the device. SnapAction stores resource cards locally with SwiftData, though AI analysis requires backend processing.
Why does On-device processing matter?
On-device processing matters because sensitive inputs can be analyzed locally, which can reduce latency and avoid unnecessary server upload.
Related pages
Next step
Turn iPhone screenshots into action-ready resource cards with AI.
View on GitHub