Memoria
Another way back to a photo.
Screens
Project concept · work in progress
What it does
Make a personal photo collection easier to browse, relate and organize through local metadata and image analysis.
How it works
- Step 1Choose a library
The user selects a library source to index.
- Step 2Build the index
Metadata and thumbnail services make large originals easier to browse.
- Step 3Find connections
Image features support similarity, duplicate candidates and face-related grouping.
- Step 4Review candidates
The user examines timelines, people, maps and candidate groups before acting.
Photo libraryFiles · metadata
Read selected images and their available metadata.
Library workspaceSwiftUI · AppKit
Explore photos, maps and grouping candidates.
Image analysisApple Vision
Detect faces and compute supported image features.
Similarity checksSHA-256 · perceptual hashes
Separate byte-identical files from visually similar candidates.
Local indexSQLite · ImageIO
Keep extracted metadata and analysis results locally.
Optional face modelONNX Runtime · ArcFace
Use the optional model when present and retain backend provenance.
Stack
SwiftUI · AppKitSQLite3ImageIO · QuickLookApple VisionCryptoKit SHA-256Perceptual image hashesOptional ONNX Runtime · ArcFaceMapKit
Local photo exploration. Similarity is a candidate signal, not proof of identity.