WeBuyBack

Inspiration

WeBuyBack was inspired by the frustration of trying to sell used items on traditional platforms—cluttered interfaces, spam messages, and slow listing processes. We asked: What if selling something locally was as easy as talking? That single question sparked the idea: a voice-powered resale platform built for speed, simplicity, and trust.

What it does

WeBuyBack allows users to list used items for sale using just their voice. Instead of typing out a full description, users can record a quick voice note, and our system transcribes and posts it instantly. Buyers nearby can browse listings, contact sellers, and arrange local pickups—all through a clean, user-friendly interface. We're blending voice tech with local commerce to make secondhand selling faster and more accessible.

How we built it

We built WeBuyBack using the following stack:

• Frontend: Next.js and Tailwind CSS for a fast, responsive UI
• Backend: Supabase for real-time data, authentication, and storage
• Voice Transcription: Deepgram API for converting voice recordings into clean text
• Payments: Stripe for gating premium features and handling subscriptions
• Guest Access: Custom ID + PIN login system for temporary users without full accounts

We focused heavily on clean UX, minimal steps to list, and reliable voice transcription. We integrated Deepgram for speech-to-text and connected it directly with our Supabase backend to store listings and transcriptions in real time.

Challenges we ran into

• Voice Transcription Accuracy: Making sure Deepgram handled different accents, background noise, and varying speech speeds was tricky.
• Guest Session Handling: Implementing a secure, frictionless way for guests to use voice features without signing up took some careful logic with ID + PIN handling.
• Gating Features: Stripe integration wasn’t just about payment—it had to also smartly toggle premium features in the UI without breaking the guest/user experience.
• Latency & Load Times: Ensuring fast uploads and real-time syncing between audio, transcription, and listing updates was more complex than expected.

Accomplishments that we're proud of

• Fully functional voice-to-listing feature with real-time transcription
• Guest user system that works seamlessly without requiring full sign-ups
• Clean UI that’s intuitive and responsive across devices
• Built an entire resale platform in a short timeframe with integrated AI and payment systems

What we learned

• How to integrate real-time voice transcription into a live marketplace
• The value of keeping UX simple—voice input can remove huge friction from traditional forms
• How to work with API rate limits, guest authentication logic, and payment gating
• Designing for trust and safety in a local selling environment is critical, especially for first-time users

What's next for WeBuyBack

• Voice Credits System: Instead of subscriptions alone, users will earn or buy voice credits to list items
• AI Listing Enhancer: Auto-generate better titles, categories, and prices using AI from voice descriptions
• Featured Listings: Let users pay to boost their listings locally
• Analytics Dashboard: Show sellers insights like average prices, trending items, and buyer demand
• Mobile App Launch: Extend the experience beyond web with push notifications and camera integration
• Local Business Partnerships: Connect with repair shops, movers, or thrift stores to create win-win affiliate systems

WeBuyBack is just getting started, but we're already rethinking what local resale should feel like in 2025; fast, personal, and voice-first.

Built With

• cohere
• deepgram
• lovable
• supabase
• vercel