What it is
Photographer-facing culling with calibrated yes/no labels.
What it does: Scores each photo against label questions and optional face matches, then lets you tune thresholds without re-running.
Why it belongs in the showcase: Substantial native app that treats Jev as a decision layer, not a captioner.
Caveats: Upload privacy; beta images; desktop-only.
Evidence
Public MIT desktop app (0★ at intake, size ~37931) pushed through 2026-10-16. README documents built-in labels (Blurry, Smiling, Eyes closed, Over/Underexposed), per-folder custom filters, TypeSafe Jev vs OpenAI switch, jev_client.py System One HTTP client, and notes that Jev images are Dreamer-ready beta. Hungarian README also present.
Caveats
- Photos leave the machine when using hosted Jev/OpenAI
- Image features are beta on TypeSafe
- Desktop packaging; not a web SaaS
Unofficial case page. Stars / list position observed for heat; do not treat heat as a quality SLA. Review the upstream README before you depend on it.