TasteStack

One app. AI recipes, real store prices, and voice-guided cooking.

Comment

Inspiration

We built TasteStack for people like us: college freshmen, new grads, and anyone who's just learning to cook. We were tired of bouncing between recipe sites, guessing at grocery costs, and fumbling with our phones over a hot stove. We wanted one app that gets you from "I'm hungry" to "I made that" without the overwhelm. TasteStack is the kitchen wingman we wished we had on day one.

Problem Statement

Learning to cook on your own is hard. Budgets are tight. You don't know what things cost at real stores. Recipe sites assume you already know the basics. Staring at a screen with flour on your hands doesn't work. New cooks and students need one place that gives them real recipes, real prices, and real-time voice guidance so they can cook with confidence instead of stress.

What it does

  • AI recipes – Tell the app what you're craving or what you have. Get a full recipe with clear ingredients and steps, no cooking degree required.
  • Real store prices – See what ingredients actually cost at Kroger, Walmart, and Harris Teeter. Shop where it's cheapest and stay on budget.
  • Voice-guided cooking – Hands-free step-by-step narration so you can cook without touching your phone. Perfect when you're learning and your hands are busy.
  • Pantry and stores – Track what you have, find stores near you, and check what's in stock. Less waste, fewer surprise trips.

How we built it

  • Frontend: React, TypeScript, Vite, shadcn-ui, and Tailwind CSS for a fast, clear interface.
  • Backend: Supabase for auth and data; Edge Functions power the AI (Gemini) and talk to grocery APIs.
  • Voice: ElevenLabs and Gemini work together so each cooking step is read aloud clearly.
  • Stores: We plug into Kroger, Walmart, and Harris Teeter so price comparison is real, not estimated.

Challenges we ran into

  • Getting the AI to return recipes in a consistent format so the app can show ingredients and steps reliably.
  • Dealing with rate limits and flaky responses from multiple grocery APIs.
  • Making voice narration stay in sync with recipe steps so the experience feels natural when you're actually cooking.

Accomplishments we're proud of

  • An app built for real beginners: no jargon, no assumptions, one clear path from idea to plate.
  • Real prices from real stores so students and new grads can actually plan around their budget.
  • Voice-guided cooking that works when your hands are full and you're still learning.
  • One app that replaces scattered tabs, guesswork, and "how do I even start?" for people learning to cook.

What we learned

  • How to design AI prompts so recipe and cost-saving outputs are reliable and easy to use in the app.
  • How to integrate and fall back gracefully when multiple external APIs are involved.
  • Why voice and minimal taps matter: learning to cook means messy hands and full attention on the food.

What's next for TasteStack

  • More stores and regions so more students and new cooks can see real prices where they shop.
  • Meal planning and "use what you have" suggestions so beginners can cook more and waste less.
  • Smarter lists and pickup/delivery options that fit how students actually shop.
  • Better pantry tracking and expiry nudges so learning to cook doesn't mean throwing money away.

Built With

Share this project:

Updates