🎵 Inspiration
Music has always been a gateway to imagination. We've all had that moment where a song transports us to another world - whether it's the epic journey in a power ballad or the intimate story in a folk song. We realized that while music streaming gives us access to millions of songs, we're still passive listeners. What if we could step inside the music and become part of the story?
The inspiration came from watching how different people interpret the same song completely differently. A child might hear an adventure story where an adult hears a love ballad. We wanted to create a platform that honors these personal interpretations while making them interactive and age-appropriate.
🚀 What it does
Lyric Odyssey transforms any song into a personalized, interactive narrative adventure. Users input song lyrics, and our AI creates a unique story where they become the protagonist. The experience includes:
- Age-adaptive storytelling that adjusts complexity, vocabulary, and themes
- Interactive choices that branch the narrative in different directions
- AI narrator integration with Tavus for conversational storytelling
- Voice commands for hands-free interaction
- Story archiving to save and share completed adventures
- Each journey is unique - the same song can generate completely different adventures based on user choices and age settings.
🛠 How we built it
- Frontend: React with TypeScript, styled with Tailwind CSS for a modern, responsive interface
- AI Integration: Google Gemini 2.0 Flash for narrative generation with carefully crafted prompts
Voice Technology:
- Web Speech API for voice commands and recognition Tavus AI for conversational narrator experiences
- Architecture: Modular component design with React Context for state management Deployment: Vite for development with Supabase backend integration
- Edge Functions: Supabase functions for server-side AI processing and image generation
The key innovation was developing age-appropriate prompt engineering that dynamically adjusts story complexity, vocabulary, and themes based on user age while maintaining narrative coherence.
🚧 Challenges we ran into
AI Prompt Engineering: Creating prompts that consistently generate age-appropriate, engaging narratives while maintaining story coherence across multiple scenes was incredibly complex. We had to balance creativity with safety.
Real-time Voice Integration: Implementing seamless voice commands and conversational AI required careful coordination between multiple APIs and handling various browser compatibility issues.
Age-Appropriate Content: Ensuring content remains appropriate across all age groups (5-120) while still being engaging required extensive testing and refinement of our AI guidelines.
Performance Optimization: Managing multiple AI API calls, voice processing, and real-time interactions without compromising user experience required careful optimization.
Cross-browser Compatibility: Voice recognition and media APIs work differently across browsers, requiring robust fallback systems.
🏆 Accomplishments that we're proud of
Seamless Age Adaptation: We successfully created an AI system that dynamically adjusts storytelling complexity, vocabulary, and themes for any age group while maintaining narrative quality.
Multi-modal Interaction: Integrating text, voice, and conversational AI into a cohesive experience that feels natural and engaging.
Innovative Use of Music: We're the first platform to transform song lyrics into interactive, personalized adventures with AI narration.
Production-Ready Design: Created a beautiful, intuitive interface that feels polished and professional, not like a typical hackathon project.
Scalable Architecture: Built with modular components and clean separation of concerns, making it easy to extend and maintain.
📚 What we learned
AI Prompt Design is an Art: Crafting prompts that consistently produce desired outputs across different scenarios requires deep understanding of model behavior and extensive iteration.
Voice UX is Complex: Designing voice interactions that feel natural requires considering context, error handling, and user expectations in ways traditional UI doesn't.
Age-Appropriate AI is Challenging: Creating AI that adapts to different age groups isn't just about vocabulary - it's about understanding developmental psychology and storytelling techniques.
Integration Complexity: Combining multiple AI services (Gemini, Tavus, Web Speech API) requires careful orchestration and robust error handling.
User Experience First: The most sophisticated AI is worthless if the user experience isn't intuitive and delightful.
🔮 What's next for Lyric Odyssey
Enhanced Personalization: Machine learning to remember user preferences and create increasingly personalized adventures based on past choices and favorite genres.
Collaborative Storytelling: Multi-player modes where friends can experience adventures together, making choices collectively.
Music Integration: Direct integration with Spotify, Apple Music, and other platforms for seamless song import and playlist-based adventure campaigns.
Educational Features: Curriculum-aligned content for schools, with lesson plans that use music to teach creative writing, literature, and storytelling.
Creator Tools: Allow musicians and storytellers to create guided adventure templates for their songs, opening new revenue streams.
Mobile App: Native iOS and Android apps with offline capabilities and enhanced voice interaction.
Community Features: User-generated content sharing, adventure ratings, and a marketplace for premium story experiences.
Advanced AI Narrators: Custom voice training for personalized narrator personalities and celebrity voice partnerships.
Log in or sign up for Devpost to join the conversation.