AI Hotdog Detector 2.0

An AI-powered app that not only detects whether you have a hotdog or not but also creates a hilarious, personalized backstory for it—one so imaginative that even you, the buyer, won’t believe it!

Comment 3

Inspiration

Ever watched the Silicon Valley series? If not go please! The idea of this web application comes from the inconic "Not a Hotdog " app that Jinyang built, which the app only does 2 things, object detection and outputs a response of whether it's a hotdog or not. While the app that Jinyang built was brain rotting enough. I came to a thought why not have a story behind it? What can be more brain rotting than having a Gen AI talk about a random hotdog's adventure to you? Which this app is for you! Who wants to know about HOTDOGS!

What it does

Unlike Jinyang's binary hotdog detector, AI Hotdog Detector 2.0 is a full-fledged hotdog investigation system:

  • Upload any suspicious image, and our AI will determine if it's truly a hotdog or just another imposter trying to infiltrate the wiener world
  • Each detection comes with a unique AI-generated story about your subject's life journey (ever wondered about a hotdog's origin story? Now you'll know!)
  • Track your hotdog detective career with achievements like "Mustard Master" when you catch real hotdogs and "Imposter Hunter" when you expose the frauds
  • Build your own Hall of Fame with a profile that showcases your greatest detective work and most memorable cases

How we built it

Used somewhat meta stacks:

  • Next.js 14+ for the frontend smooth experience
  • Supabase for authentication and database (storing all those juicy hotdog stories)
  • Google Vision API for the awesome hotdog detection
  • OpenAI for generating those dramatic hotdog backstories
  • Tailwind CSS + shadcn/ui to make everything look as aligned as possible

Challenges we ran into

  • Setting up Google OAuth was terrible not to speak the original OAuth
  • Dealing with 'Duplicate Key' Errors (Too many same hotdogs)
  • Trying to stay focus and code with a numerous Hotdog images

Accomplishments that we're proud of

  • Successfully built an app which is way more brain rotting than the original
  • Built a proper user authentication system that actually worked
  • Made hotdog detection both fun and slightly educational (maybe?)

What we learned

  • Google Vision API can be pickier about hotdogs than a food critic
  • Supabase makes database management slightly easier (but OAuth was a disaster)

What's next for AI Hotdog Detector 2.0

  • Implement real-time hotdog detection using device cameras
  • Create a global leaderboard for top hotdog detections (Why would you do that?)

Built With

  • googlevisionapi
  • next.js
  • openai
  • react
  • supabase
Share this project:

Updates