VoiceBio

Based on the VoiceBio project code I've analyzed, here's the submission writeup:

Inspiration

The inspiration for VoiceBio came from recognizing that everyone has a unique story worth preserving, but traditional biography writing is time-consuming and intimidating. We wanted to democratize storytelling by making it as simple as having a conversation. With aging populations and the importance of preserving family histories, we saw an opportunity to use AI and voice technology to capture life stories in a natural, engaging way.

What it does

VoiceBio is an AI-powered platform that transforms personal stories into beautiful biographies through voice interviews. Users participate in guided conversations across six life categories (Early Life, Education & Career, Major Life Events, Personal Relationships, Achievements, and Hobbies). The AI conducts natural follow-up questions, transcribes responses in real-time, and generates compelling biographical content. Users can then export their stories as professionally styled PDFs with customizable templates, colors, and fonts, or listen to AI-generated audiobook versions.

How we built it

We built VoiceBio using a modern React/TypeScript frontend with Tailwind CSS for styling. The backend leverages Supabase for authentication and data storage, storing interview sessions and biographical content. We integrated ElevenLabs for text-to-speech functionality, providing a natural female voice for questions and audiobook generation. OpenRouter API powers the AI content generation, creating personalized biographies from interview responses. The PDF generation system uses custom styling templates to match the on-screen book view, ensuring WYSIWYG functionality.

Challenges we ran into

One major challenge was achieving perfect synchronization between the on-screen book view and PDF output - ensuring fonts, colors, and layouts matched exactly. We also faced complexity in managing the conversational flow, balancing AI-generated follow-up questions with structured interview progression. Real-time speech transcription accuracy required careful error handling and user editing capabilities. Additionally, integrating multiple AI APIs (ElevenLabs, OpenRouter) while maintaining smooth user experience and handling API rate limits proved challenging.

Accomplishments that we're proud of

We're proud of creating a seamless voice-to-biography pipeline that feels natural and intuitive. The AI interviewer maintains engaging conversations while systematically gathering comprehensive life stories. Our WYSIWYG book view with multiple professional styling templates gives users beautiful, publication-ready results. The real-time transcription with editing capabilities ensures accuracy, and the audiobook generation adds a personal touch. The responsive design works across devices, making biography creation accessible to users of all technical backgrounds.

What we learned

We learned the importance of user experience design in AI applications - making complex technology feel simple and approachable. Working with multiple AI APIs taught us about handling asynchronous operations, error management, and graceful degradation. We discovered the value of incremental progress tracking and allowing users to resume interviews, as biography creation is an emotional and time-intensive process. The project reinforced the power of combining multiple AI technologies to create something greater than the sum of its parts.

What's next for VoiceBio

Next, we plan to add collaborative features allowing family members to contribute to biographies together. We're exploring multi-language support for global accessibility and considering integration with genealogy platforms. Advanced AI features like automatic photo organization and timeline generation are in development. We also want to add video recording capabilities and expand export formats (ePub, interactive web pages). Long-term, we envision VoiceBio becoming a comprehensive family history platform with sharing capabilities and legacy preservation tools.

Built With

• elevenlabs
• postgresql
• react
• supabase
• typescript