Landing Page
Attendee Event Selector
Attendee Anonymized
Attendee Camera Anonymizer
Organizer Photo Intake
Organizer Privacy List
Organizer Security Audit Log

Anonify: Privacy-First Event Photo Redaction

💡 Inspiration

We’ve all been there: you attend a wedding, a tech conference, or a party, and suddenly your face is in a hundred photos shared on a public Google Drive or social gallery. For many, this is a privacy nightmare.

We wanted to build a tool that gives power back to the attendee without making the organizer’s life impossible. Anonify was born from the idea that “opt-out” should actually mean something in the digital age.

🛠️ How We Built It

We built Anonify as a multi-layered, local-first system to ensure maximum privacy:

Frontend: Next.js with TypeScript and Tailwind CSS for a sleek, responsive organizer dashboard and a simple, mobile-friendly attendee “Opt-Out” portal
Auth: Clerk for secure organizer authentication
Backend: Flask API managing relational data with SQLite (portable, self-contained database)
AI / Vision: Python microservice using OpenCV and face recognition
AI Usage: Claude, Codex, Gemini

The Flow

Attendees join an event via a unique HUSKY-KEY
If they opt out, they take a quick selfie (reference image)
The selfie is stored as a Base64 string in SQLite
Organizers upload event photos
AI matches faces in photos against opt-out selfies
Organizer reviews detections and clicks “Anonymize” to apply persistent blurs

🧠 What We Learned

We learned that handling facial recognition requires a careful balance between user experience and privacy.

A major focus was confidence scoring—instead of making risky assumptions, the system flags uncertain matches for manual review.

We also learned how to effectively coordinate a Next.js frontend with a Python AI backend in near real-time.

🚀 Challenges We Faced

🔐 The “Turnstile” Trouble

Cloudflare Turnstile bot detection interfered with our local Clerk authentication flow. We had to debug middleware and browser security headers to stabilize development.