Vision Tags

Problem

Filtering media by filename alone is inadequate. Users need content-aware filtering — finding all photos with faces, all dark shots, or all images containing text.

Solution

FlexGrid runs 18+ Vision framework classifiers in a background sweep: faces, bodies, text, animals, scenes, brightness (dark/normal/bright), B&W/monochrome, horizon tilt, blur/smudge, letterbox/pillarbox, composite/collage detection, barcodes, and feature print similarity. Each tag has an SF Symbol icon. Filter the grid to any tag with one click.

Tier Availability

• Free: Not available.
• Plus / Pro: Full vision tagging with all 18+ classifiers and one-click filtering.

UI Flow

1. Load media. A background sweep runs automatically (progress shown in status bar).
2. Open the Vision Tags filter from the menu or toolbar.
3. Select a tag to filter the grid to matching items only.
4. An ActiveFilterPill shows the current filter and can be dismissed.

Known Limitations

• Smudge/lens detection requires macOS 26+.
• Background sweep can take minutes on large libraries (progress is shown).
• Classification accuracy depends on Vision framework; edge cases may be mis-tagged.

Marketing One-Liner

18+ AI-powered content tags — faces, text, animals, brightness, blur, and more — let you filter by what’s in the image, not just its filename.