Live pipeline · open playground
One upload. Four stages. Two finished products.
Upload a portrait photo. The pipeline runs face-landmark detection in your browser, runs a vision-language rule gate against a default school rule list via OpenRouter, runs identity-preserving diffusion enhancement on Replicate, and composes the yearbook portrait and the ID-card photo from the same Stage-1 face-size measurement. This page uses a generic rule list. For a per-school flow with an invitation code, go to the school portal.
Age range of the subject
Drives the IPD (inter-pupillary distance) used for output scaling. Newborns have a relatively larger IPD per face size; from ~6 months onward the proportion matches the adult standard.
Adult: 1″ face → 82 px IPD · 1.25″ face → 102 px IPD
- Stage 01Geometric validation
- Stage 02School-rule gate
- Stage 03Identity-preserving enhancement
- Stage 04Dual-output composition
Upload a portrait photo
Drag and drop, or click to browse. The photo never leaves your browser except for the rule-gate call to the vision-language model.
JPG · PNG · WebP · up to ~8MB
Or try a sample photo:
No state is persisted on this page. To save submissions to a school roster, sign in to the school portal and use an invitation-code upload.