Inspiration
Stories passed down through families and communities often live only in memory. As generations move, languages fade, and elders pass away, these oral histories quietly disappear. Inspired by the theme of Identity, we are exploring the identity that stored in these stories; the way people speak, remember and narrate their lives
What it does
Family Stories allows users to record spoken stories, such as, family memories, cultural anecdotes, or personal experiences, and automatically converts them into text. The app then lets others interact with these stories through a conversational AI, asking questions, clarifying details, or exploring themes.
In short, the application:
- Records oral stories
- Convert speech to text
- Chat with the story to explore meaning, memories, and identity
How we built it
We built Family Stories as a web application using:
- Next.js and React for the frontend
- MongoDB for the backend
- Google Speech-to-text APIs to transcribe recorded audio
- Gemini API to enable conversational Q&A grounded in the original story
Challenges we ran into
Preserving authenticity: Ensuring the AI responds using the storyteller’s voice and style, without distorting meaning. Audio handling: Managing audio formats, transcription accuracy, and latency within a short hackathon timeline. Scope control: Balancing ambitious ideas with what was realistically achievable in limited time.
Accomplishments that we're proud of
- Turning a deeply human idea into a working technical prototype
- Designing an AI interaction that supports stories instead of overshadowing them
What we learned
- AI is most powerful when it amplifies human voices rather than replacing them
- Identity-focused tech requires empathy as much as engineering
What's next for Family Stories
- Multi-language support to preserve stories in native languages
- Private family collections and sharing controls
Built With
- gemini
- google-web-speech-api
- next.js
Log in or sign up for Devpost to join the conversation.