AI · App product
MyEventPics
AI · App Product · Solo · On-device FaceNet SaaS
Event photo delivery SaaS — guests scan a QR, take a selfie, and get only their AI-matched photos in seconds.
AI Photo SaaS: QR → selfie → your album
Intelligence in the app
AI meets product UX
Phone screens for the consumer loop — AI panels for the model that powers them.
AI + product story
Overview
A dual-role Android product where AI is the product: photographers create events and upload; guests find themselves with on-device FaceNet matching instead of scrolling thousands of images.
What I built
- Photographer: create events, QR / App Links, bulk upload with foreground service
- Guest: scan QR → Event ID → liveness selfie → FaceNet match → download
- Google Drive storage for media; Firestore for events, plans, and embeddings
- On-device ML Kit + FaceNet TFLite (+ optional Python FastAPI face API)
- Subscriptions with usage caps; Razorpay payments + admin verification
- React admin: photographers, events, revenue, service status
- Public Next.js guest/payment site on Vercel
What's unique
- Privacy-first: originals stay in the photographer’s Google Drive — Firebase keeps metadata + embeddings only
- On-device FaceNet matching with liveness selfie (blink/turn) so guests get ‘your photos only’
- Photographer SaaS loop + viral guest QR flow in one Android product
Why it matters
- Ends WhatsApp / Drive dump chaos for wedding and party photographers
- Guests leave with their album in seconds — not after hours of searching
- Clear monetization with plan limits, Razorpay, and an admin console
Craft
How AI ships inside the app
AndroidJavaCameraXFaceNetML KitFirebaseGoogle Drive APIFastAPIReactNext.jsRazorpay
Try the product