frameSeek — A Local Image Archive
Search Dropbox, disks and NAS volumes as one photographic memory—without moving the originals or sending a single image to an AI service.
frameSeek creates one searchable index across Dropbox, local folders, external SSDs and mounted NAS shares while every original remains exactly where it is. EXIF, IPTC and XMP metadata sit beside local previews, so the archive remains searchable even when a drive is disconnected.
Keyword search can be combined with natural-language similarity search powered entirely on the Mac by multilingual SigLIP 2 Core ML models. Saved filter recipes mix tags, semantic phrases and dragged-in reference images, while equipment, projects and resumable exports turn search results into practical working sets.
Analogue photography is part of the data model rather than an afterthought. Scans can be linked to physical negative sheets and storage locations, collected into lab lists and printed as A4 contact sheets. The interface changes between analogue and digital contexts so specialist tools appear only where they make sense.
Duplicate handling is deliberately cautious. A separate perceptual fingerprint identifies genuinely equivalent files, recommends the strongest master and can merge tags and project associations before anything is deleted. Write tests, explicit confirmations and XMP updates to the original files protect an archive whose metadata and negatives may be irreplaceable.
