Privacy First
Local Processing
Face detection runs locally. Embeddings stay on disk; nothing is sent to a cloud service unless you configure an external endpoint.
Visual Intelligence
Face Recognition & People Grouping
Cluster faces with local ONNX models. Name people once and browse photos by person across your archive without cloud upload.
Key Features
What each module does.
Native Windows app. Hashing, indexing, and batch apply run on your machine.
Accuracy
Recognize People Over Time
Identify family and friends across years of photos, different ages, lighting, and angles, all processed locally.
Organization
Auto-Cluster Groups
Automatically group thousands of scattered photos into dedicated 'People' collections instantly.
The Advantage
Photos grouped by person
Name a face once, then open every photo that person appears in across the library.
Visual Duplicate Detection
Find near-identical shots of the same person to keep only the best frame from a burst.
Searchable People Collections
Name people once and find their photos across your library, in Smart Tags, and in Universal Search.
Comparison
DupeZappa vs PhotoPrism
| Feature | DupeZappa | PhotoPrism |
|---|---|---|
| Deployment | Windows desktop app; no Docker required | Self-hosted server with Docker and web UI |
| Library model | Faces on existing folder paths; no import pipeline | Indexed photo library with import and originals |
| Face pipeline | Local ONNX models (YuNet + AuraFace) on your PC | Server-side ML stack in the PhotoPrism instance |
| Browse UX | People clusters, Smart Tags, Universal Search | Albums, places, calendar, labels, web gallery |
| Beyond faces | Duplicates, junk, search, rename, preview-first undo | Full DAM workflows and browsing |
| Price | One-time purchase (early access) | Open source + your own server and hosting cost |
Also compare: vs Immich
Take Control
Preview first, then apply in batches.
Run scans locally, review the results, and apply with Recycle Bin and operation history.