Zesty

The Unrecommendation Engine

Comment

Zesty – The Unrecommendation Engine

🪄 Inspiration

We live in a time where every app recommends the same kind of content.
If you watch one romantic movie, your feed gets filled with more romance.
If you listen to mellow songs, you rarely see energetic music suggestions.

This creates a bubble around our tastes — a comfort zone that limits discovery.

We started Zesty to break that bubble.
We wanted to help people discover things different from their usual likes.
Not just more of the same, but something unexpected and exciting.

That’s where the idea of the "Unrecommendation Engine" came from.

🎯 What It Does

Zesty helps people find content that is completely opposite to what they usually enjoy.
Whether it’s music, movies, books, or food, Zesty uses cultural data and AI to suggest something fresh and different.

  1. Taste Journey – Users answer simple questions about what they like.
  2. Opposite Match – Zesty uses the Qloo API to find the farthest possible point from their current preferences.
  3. Beyond Suggestions – Daily reflection prompts, comfort-zone challenges, and taste evolution tracking.
  4. Taste Nemesis – Meet someone who loves what you hate, and share your journeys.

🛠 How We Built It

  • Frontend: React + Tailwind CSS, powered by Next.js App Router for speed and simplicity.
  • Backend: Supabase for database, authentication, and edge functions.
  • AI Integration: Gemini 1.5 Pro model for generating prompts, reflections, and taste insights.
  • Cultural Data: Qloo API for intelligent, data-backed recommendations.

We focused on keeping the app light, fast, and intuitive with meaningful suggestions.

⚡ Challenges We Faced

  • Defining "opposite" tastes without making them feel random.
  • Designing short but effective onboarding to capture preferences.
  • Balancing user comfort with new experiences.
  • Smoothly connecting Qloo and Gemini.
  • Keeping the UI simple while offering depth.

🏆 Accomplishments

  • Created the concept of an "Unrecommendation Engine".
  • Integrated Qloo, Gemini, and Supabase into one working app.
  • Designed a playful yet smooth user experience.
  • Encouraged people to explore outside their usual choices.

📚 What We Learned

  • People are open to trying new things if guided the right way.
  • AI can be used to broaden tastes, not just reinforce them.
  • Cultural data becomes powerful when paired with AI.
  • Exploration should be guided, not forced.

🚀 What's Next for Zesty

  • Mobile App for better accessibility and UX.
  • Streaks & Levels to gamify taste discovery.
  • Partnerships with Spotify, Goodreads, Netflix for deeper integrations.
  • Creator Mode to let users design their own taste challenges.
  • Exploring use cases in education, team-building, and diversity training.

Built With

Share this project:

Updates