Object Talk

Bring everyday objects to life with interactive AI personalities, multi-object debates, and instant web-to-app subscriptions

Comment

Inspiration

After the sudden demise of my father, I had a deep realization that I could no longer remain passive or unproductive. I had to prove something to myself. My academic journey was long and non-traditional—it took me seven years to complete my graduation in Computer Science & Engineering, and even after five years post-college, I remained unemployed. I desperately wanted to rebuild myself into a better version of who I could be.

One day, while brainstorming ways to reignite my journey, my brother asked a whimsical question: "What if the everyday objects around us could actually talk?" That spark led to my very first prototype. Shortly after, I stumbled upon the RevenueCat Ship-a-ton hackathon on X (Twitter). Seeing that gave me the exact push, accountability, and deadline I needed to take this project seriously, push beyond my self-doubt, and build a real production-grade mobile app.

What it does

Object Talk turns everyday physical objects into living, speaking personalities using multimodal AI:

  • Point & Chat: Snap a photo of anything—a coffee cup, your sneakers, a tired laptop, or a houseplant—and the AI brings its voice to life with expressive character dialogue and audio.
  • Dynamic Personas & Savage Roasts: Choose how items talk back! From cheeky banter and sharp, sarcastic roasts to philosophical musings and warm encouragement.
  • Origin Lens: An educational discovery mode that reveals where items physically came from—tracing their raw materials, global supply chain, and historical manufacturing journey.
  • Multi-Object Debates: Put two to four items together (e.g. coffee vs. tea) and listen to them playfully argue with each other.
  • Memory Library & Freemium Economy: Save favorite conversations and unlocked cards. Powered by RevenueCat, users can unlock unlimited scans, secret personas, and extra daily credits via Day Passes or flexible Pro subscriptions.

How we built it

While I hold a computer science degree, I wasn't fluent in coding complex modern apps from scratch. However, I understood core architecture, how tools connect, how to read documentation, and how to debug iteratively.

Leveraging modern AI tools like Gemini and Claude as my pair-programmers, I acted as the product manager and lead architect—directing implementation, testing every build on physical Android devices, and steering the app to Google Play production:

  • Frontend: Built with React Native & Expo SDK 54, utilizing Expo Router, React Native Reanimated, and Skia for a fluid camera HUD.
  • Multimodal AI & Speech: Integrated vision AI models to identify items, generate persona dialogue, and stream realistic text-to-speech audio.
  • Monetization Engine (RevenueCat): Configured a multi-tier economy featuring consumable Day Passes (120 coins) and auto-renewing subscriptions (1-month, 3-month popular, 6-month). RevenueCat seamlessly synchronizes entitlements across our mobile client, Google Play Billing, and backend database.
  • Backend & Security: Hosted on Render with Supabase PostgreSQL, utilizing anonymous hardware-backed JWT authentication (Expo SecureStore) and zero-login onboarding.

Challenges we ran into

  • Navigating the Syntax Barrier: Directing AI to write high-performance React Native code required rigorous architectural judgment, linting, and catching edge-case bugs.
  • Google Play & RevenueCat Pipeline: Setting up Google Cloud Service Accounts, configuring in-app billing in Play Console, mapping products in RevenueCat, and passing stringent closed testing requirements was a massive learning curve.
  • Subscription Lifecycle Edge Cases: Handling real-world payment states—ensuring coins are credited instantly, handling grace periods correctly when a subscription is canceled without prematurely revoking access, and providing offline fallback states.

Accomplishments that we're proud of

  • From Unemployed to Published Developer: Successfully completing closed testing, passing Google's production review, and launching Object Talk live in Production worldwide on the Google Play Store.
  • Seamless In-App Monetization: Building a real, fully functioning RevenueCat paywall with working subscriptions and consumable purchases.
  • Personal Growth: Overcoming years of hesitation and proving to myself that with relentless focus and modern tools, I can build and ship world-class software.

What we learned

I learned that being a developer in 2026 isn't about memorizing syntax; it's about resourcefulness, understanding documentation, breaking down complex systems, and having the grit to finish what you start. Building Object Talk restored my confidence and proved that delayed beginnings do not define the end of your story.

What's next for Object Talk

Launching on the Play Store is just Day 1. Here is our roadmap:

  • Next-Level Roast Quality: Fine-tuning custom prompt chains and comedic timing to make roasts punchier, culturally hyper-relevant, and genuinely hilarious based on specific object wear-and-tear (e.g. noticing a cracked phone screen or scuffed shoes).
  • Viral Content Marketing: Launching dedicated TikTok, Instagram Reels, and YouTube Shorts channels showcasing hilarious object conversations (e.g., "What your gym shaker bottle says about you behind your back"). This will be our primary organic user acquisition engine.
  • Interactive Live Voice (Full Duplex): Transitioning from turn-based audio playback to real-time conversational voice streaming, so users can have continuous back-and-forth verbal banter with their objects.
  • Expanding Distribution: Bringing Object Talk to alternative Android app stores like Indus Appstore to capture regional audiences across emerging markets.

What we learned

What's next for Object Talk

Built With

Share this project:

Updates

Submission history