Faike Buster

AI-powered product scanner app that verifies authenticity of products and flags fake(counterfeit) items.

Comment

Inspiration

Counterfeit products cost consumers billions annually. After purchasing many fake products myself, I wanted to build an AI tool that instantly verifies authenticity using computer vision and crowd-sourced data—like a "Shazam for fakes." I was looking to buy a popular skincare product — Anua Azelaic Acid — but I wasn’t sure if the one I found was original or counterfeit. I discovered that many people on platforms like TikTok had the same issue. Some were unknowingly using fake skincare products that caused bad reactions or didn’t work. There wasn’t a quick and reliable way to confirm a product’s authenticity in real time.

That’s when I realized: what if I could scan a product, and AI could instantly tell me if it was fake or real — like a smart assistant in my pocket?

That’s how FaikeBuster was born — a web app that lets users scan products and uses AI to help detect counterfeits in real-time.

What it does

FalkeBuster scans product packaging and: Compares it to genuine products using GPT-4 Vision and Google Vision Cross-checks Reddit/forums for counterfeit reports Generates a confidence score (e.g., "81% authentic") with actionable feedback

How we built it

Backend: Node JS, OpenAI API + Google Vision + SerpAPI (for web scraping) with Bolt.new Frontend:Next JS with Bolt.new Database: Supabase (stores user-submitted fake/real examples) Hosting: Netlify

Challenges we ran into

Rate limits: OpenAI and Google API quotas required optimizations (batching, retry logic)/billing issue-no money. I ran into configuration issues with ElevenLabs, like API key restrictions. Scraping restrictions: Reddit/TikTok APIs forced us to use SERP workarounds. It was also a challenge to structure the AI prompt logic so the app knew how to flag only relevant products

Accomplishments that we're proud of

Created a real-world AI use case that helps people avoid harmful counterfeit beauty and healthcare products Designed a clean and intuitive frontend experience with Bolt.new — ready to be used by real consumers.

What we learned

The importance of building user-friendly apps for real-life everyday problems AI + human input beats pure automation (hence the hybrid approach). How to build voice experiences with ElevenLabs SDK Learnt how to manage environment variables, connect API keys, connect to Supabase and deploy on netlify

What's next for Faike Buster

Launch a mobile version of the app to make scanning products easier while shopping in physical stores. Crowdsource product reports — let users flag suspicious items, share experiences, and contribute to a global fake product database. Enhance the AI assistant to give personalized recommendations, recall past scans, and explain why a product might be fake.

Built With

  • bolt.new
  • elevenlabs
  • netlify
  • openai-api
  • serpapi
  • supabase
  • zxing
Share this project:

Updates