KORENANI - Turn the World Into Your Family Encyclopedia

Kids photograph the world, hear names in 9 languages, and build their own private encyclopedia. Learning that grows with them.

Comment

🔗 Try It Now

Inspiration

As a child, I was fascinated by a game where captured characters were saved into a personal encyclopedia. Collecting fed my curiosity, and I memorized nearly every name. I used to imagine: what if this could happen in real life?

Years later, as a parent, I saw the chance to make that dream real. Today, AI can recognize the world around us. A trip to the zoo, a walk through the city, even a casual snapshot of a train—all can be stored as part of a child's living encyclopedia. Watching my own child's passion for cars and trains, I realized how toys and books nurture curiosity. I wanted to create something that connects those sparks of wonder with real-world learning and family memories. That became Korenani.

What it does

Korenani transforms everyday moments into a multilingual learning adventure - no ads, no complexity, just pure curiosity.

Turn any family outing into an educational experience:

  • 🦁 Zoo visits become a personalized wildlife encyclopedia
  • 🚂 Train spotting turns into a transportation collection
  • 🌸 Nature walks build a living garden guide
  • 🏙️ City exploration creates an architecture album

Core Experience

Children take photos of anything interesting. Korenani analyzes the photo, speaks the name in their chosen language, and saves it to their personal collection. Each discovery becomes part of a growing family encyclopedia that makes learning feel like play.

Key Features

🎯 Instant Recognition from Photos

  • Take a photo of anything interesting
  • AI analyzes the photo for identification and naming
  • Animals, vehicles, plants, everyday objects, and more

🔬 Specialist Mode

  • Detailed analysis for flowers and insects: Not just species names, but also cultivation tips, habitat information, and other specialized knowledge
  • Supporting deeper learning: When children's interests deepen, provide more detailed knowledge
  • Generation from keywords: Add family-specific terms like "grandpa" or "grandma" to the encyclopedia from keywords, even without photos

🌍 9-Language Support

  • Clear audio playback in: English, Japanese, Spanish, French, German, Italian, Portuguese, Chinese (Mandarin), and Korean
  • Natural pronunciation using iOS native voice synthesis
  • Switch languages instantly with one tap
  • Children naturally learn pronunciation through repetition

📚 Smart Collections

  • Save discoveries to custom albums ("Zoo Day", "Our Garden", "Favorite Cars")
  • Public Albums: Curated collections (Animals, Vehicles, etc.) copyable with one tap
  • Review and replay anytime - your family's personal encyclopedia

🎮 Quiz Mode

  • Test knowledge by album and language
  • "Which one is the lion?" - interactive learning
  • Reinforces memory through playful repetition

🏆 Achievements & XP

  • Earn badges for taking photos and completing quizzes
  • Gentle motivation system encourages exploration
  • No pressure, just positive reinforcement

👶 Toddler-Friendly Design

  • Large buttons designed for small hands
  • Simple, intuitive navigation
  • Usable independently by children as young as 2-3
  • High-contrast UI works outdoors in bright sunlight

🔒 100% Ad-Free & Private

  • No advertisements or interruptions
  • Family photos remain completely private
  • No third-party data sharing
  • Public content is only KORENANI-curated albums

How we built it

Architecture

Frontend

  • Swift/SwiftUI for iOS development
  • iOS native voice synthesis (AVSpeechSynthesizer) for natural pronunciation in 9 languages
  • Large touch targets and UI patterns optimized for toddlers
  • Camera integration for photo capture and analysis
  • Offline-capable design

Backend

  • NestJS API on Google Cloud for scalable services
  • AI image recognition pipeline for photo content identification
  • Multilingual name resolution system
  • Detailed information retrieval system for Specialist Mode (species, cultivation, habitat, etc.)
  • Keyword-based content generation functionality
  • Caching layer for instant responses

Database & Auth

  • Supabase Postgres with comprehensive Row-Level Security (RLS)
  • Supabase Authentication with ownership-based access control
  • Strict storage permission rules - every photo validated against user ownership
  • Public album system with read-only access to curated content

Monetization

  • RevenueCat integration for subscription management
  • Entitlement-based feature gating
  • A/B testing capability
  • No ads, no data sales - sustainable through subscriptions

We built the app with Swift (iOS), Supabase for backend data, and integrated the RevenueCat SDK for subscription and entitlement management. This setup allows seamless handling of free trials, promo codes, and paywall logic.

Privacy-First Design Philosophy

Every architectural decision prioritizes family privacy:

  • RLS policies ensure family photos are never accessible to other users
  • Storage ownership checks validate permissions before any file access
  • Public albums are curated exclusively by KORENANI - users can copy but never share private photos
  • No tracking beyond essential app functionality
  • Local processing where possible to minimize data transmission

Design Process

  • Iterative testing with actual toddlers in various lighting conditions
  • Voice feedback design to support pre-reading children
  • Color contrast optimization for outdoor visibility
  • One-handed operation for parents holding children

Challenges we ran into

UX for Toddlers in Real-World Conditions

  • Designing flows navigable by 2-3 year olds with "wobbly fingers"
  • Ensuring buttons remain tappable outdoors in bright sunlight with glare
  • Balancing visual simplicity with feature richness
  • Creating instant feedback loops that keep young attention spans engaged

Multilingual Natural Language

  • Delivering contextually appropriate names across 9 languages
  • Handling synonyms (is it a "car" or "automobile"? "dog" or specific breed?)
  • Cultural variations in object naming conventions
  • Logic for selecting the most natural expression in each language

Specialist Mode Accuracy

  • Accurate species identification for flowers and insects
  • Retrieving and validating reliable cultivation information and habitat data
  • Appropriate handling of family-specific high-context keywords ("grandpa", "grandma", etc.)
  • Balancing specialist information with simplicity for children

Privacy at Scale

  • Implementing foolproof RLS policies that scale to millions of photos
  • Ensuring storage rules remain maintainable as features grow
  • Balancing public album discoverability with strict privacy guarantees
  • Designing a permission system parents can understand and trust

Recognition Accuracy

  • Achieving reliable identification across lighting conditions
  • Supporting diverse angles and distances children capture
  • Balancing speed (quick feedback) with accuracy
  • Quality control and error handling for recognition results

Accomplishments that we're proud of

✅ 9 languages from day one - Multilingual learning without barriers, making global education accessible to all families

✅ Uncompromising privacy - Strict RLS and storage rules protect every family. Zero compromises, zero data sales

✅ Deepening learning with Specialist Mode - Expanding learning from basic identification to specialized information as children grow

✅ Understanding family context - Can add family-specific words like "grandpa" and "grandma" to the encyclopedia from keywords

✅ Production-ready architecture - Scalable backend handling recognition, storage, and authentication with enterprise-grade security

✅ Natural learning loop - Kids replay voices simply for fun, creating organic repetition that accelerates vocabulary acquisition

✅ Ad-free experience - Sustainable through subscriptions, never through attention exploitation or data monetization

✅ Foundation for growth - Architecture ready for voice library, institutional partnerships, and Android expansion

✅ Real-world validation - My own child became the first user, and his excitement proved the concept works

What we learned

Children Learn Best from Their Own Photos When kids capture images themselves, engagement and retention skyrocket. It's not just about seeing pictures - it's about the agency of discovery.

Multilingual Playback Drives Natural Repetition Children replay the sounds simply because it's fun. This playful repetition creates language exposure without feeling like study.

Depth Matters as Children Grow At first, "That's a lion!" is enough, but as interest deepens, questions increase: "Where does it live?" "What does it eat?" Specialist Mode supports this natural growth of curiosity. Species names for flowers, cultivation tips, habitat information for insects - opening doors to deeper learning.

Family Context is Irreplaceable Every family has their own words and nicknames - "grandpa", "grandma", "doggy", etc. The ability to generate items from keywords means we can add important family members and concepts to the encyclopedia even without photos. This isn't just a convenience feature - it's a vital function for weaving family memories.

Privacy Is Non-Negotiable - As a Parent, I Cannot Compromise Building this app as a parent, privacy became paramount. Handling my own child's photos, I cannot compromise on data handling. Supabase Authentication, RLS, and storage ownership checks aren't optional features - they're essential safeguards I need as a parent. I implemented the level of privacy protection I myself want to use.

User Feedback is Everything - Especially When Your User is Right Next to You Something I always emphasize in my professional engineering work, but I was reminded again of the importance of actual user feedback. My 2-year-old can't tap accurately yet and sometimes scrolls unintentionally. This is a problem I couldn't have discovered from specs or personas. Currently planning improvements to touch targets and gesture recognition, but these discoveries only come from having users nearby and observing them use the app daily. And those smiles are the greatest motivation to keep building.

Simplicity Beats Feature Overload Early versions had too many options. The final design focuses on the core loop: see, learn, collect. Everything else is secondary.

Public Albums Solve the Cold Start Problem New users needed content immediately. Curated public albums gave families a starting point while building their own collections.

My Child Is the Best Product Manager Watching him use the app revealed insights no focus group could provide. His smiles and frustrations directly shaped every iteration.

What's next for KORENANI

Voice Library - Phase 1

Adding warmth to learning with customizable audio experiences

Currently using iOS native voice synthesis, but planning to add voice personalization features as the next step:

  • Record family voices: Parents and grandparents can record names in their own voices. Warm experience of hearing "butterfly" in grandma's voice, even when she lives far away
  • Animal sounds: Lions don't just say "lion" - they can also play real lion roars
  • Vehicle sounds: Experience actual train sounds along with the train's name
  • Environmental sounds: Make learning more multidimensional with sounds related to objects

This will evolve Korenani from a purely visual learning tool into a comprehensive sensory learning experience including auditory elements.

Enhanced Specialist Modes - Phase 2

  • Additional categories: Birds, fish, reptiles, and more specialized categories
  • Seasonal information: Flowering times for plants, insect activity periods, etc.
  • Observation records: Track the same subjects over time, recording growth and changes
  • Expert-verified content: Information verified and expanded by specialists from botanical gardens and zoos

Institutional Partnerships - Phase 3

  • Zoos & Aquariums: Custom albums for visitors to extend on-site discovery into at-home learning
  • Museums: Art and artifact collections with educational context
  • Botanical Gardens: Plant identification with seasonal updates
  • Events & Festivals: Temporary albums for special experiences

Platform Expansion - Phase 4

  • Android launch: Bring Korenani to millions more families worldwide
  • Tablet optimization: Better experience for shared family devices
  • Offline mode enhancement: Full recognition capability without connectivity

Revenue Model Evolution - Phase 5

  • Current: Subscription-based access with free tier
  • Planned:
    • Spot purchases of themed albums (e.g., "Ocean Life", "Construction Vehicles")
    • Institutional accounts enabling organizations to create branded public albums
    • Family plans for multi-device households
  • RevenueCat Growth: Advanced entitlement management, A/B testing, paywall optimization

Advanced Features - Phase 6

  • Learning paths: Guided exploration themes ("Find 10 different birds")
  • Social features: Share achievements (not photos) with extended family
  • Educator mode: Tools for teachers to create classroom collections
  • Augmented reality: Overlay additional information on real-world objects

Impact & Vision

Breaking Language Barriers in Early Childhood Education

Korenani makes multilingual learning accessible to families worldwide, regardless of their linguistic background. Parents can expose children to 9 languages naturally through daily activities - no textbooks, no classroom pressure, no expensive tutors. Just take photos, listen, and learn.

Privacy in the Age of AI

In a world where children's data is increasingly monetized and attention is commodified through ads, Korenani demonstrates that powerful AI experiences can be built without compromising family privacy. We believe learning tools should serve children, not exploit them.

Growing with the Child

Korenani isn't just a toddler app. Through Specialist Mode, it responds to curiosity that deepens from "What's this?" to "Why?" and "How?" It's a growth-oriented learning tool that children who start at age 2 can continue using at 5, 10, and beyond.

Democratizing Curiosity

Every child is naturally curious, but not every child has access to encyclopedias, educational apps, or stable internet. Korenani's simple interface and offline capabilities make learning accessible even in resource-constrained environments. The world becomes the textbook.

Building Generational Bridges

The upcoming voice library will enable grandparents to contribute to a child's learning even across continents. The keyword generation feature lets you add "grandpa" and "grandma" to the encyclopedia and record their voices. This transforms Korenani from an educational tool into a family memory keeper, preserving voices and creating intergenerational connections through shared learning.

Built With

  • swift - iOS app development with SwiftUI
  • typescript - Backend services and API
  • nest.js - Scalable backend framework
  • google-cloud - Infrastructure and AI services
  • supabase - Authentication, database (Postgres with RLS), and storage
  • revenuecat - Subscription and monetization management

How to Use (Simple 4 Steps)

  1. Grant camera access when prompted
  2. Take a photo of anything interesting
  3. Listen and learn names in 9 languages
  4. Save to collection and share with family

Built with ❤️ by a parent who believes the world is the best classroom, and that every child deserves to explore it with confidence, curiosity, and joy.

Built With

Share this project:

Updates