hackhavard

Inspiration

We’ve always been fascinated by how technology can preserve and re‑imagine history. Too often, history feels distant or abstract — trapped in books and timelines. We wanted to use AI to make it more visual, interactive, and emotionally engaging. The idea behind HistorioVision (built here as the Cultured app) is to create an experience where people can see the past come alive through dynamic visuals, spatial/AR elements, and context‑aware storytelling.

What it does

Cultured lets users capture or upload content about historical artifacts or moments, then uses AI to generate:

Contextual narratives and concise summaries
Visuals that match the historical style
Optional voice narration
An immersive AR view to explore artifacts in space

Users can browse a feed of posts, open a timeline‑like detail view, and experience the generated story and media together.

How we built it

Python (FastAPI) for backend orchestration and API management
Google Gemini for historical storytelling and summary generation
Image generation services to create historical‑style imagery
ElevenLabs for optional voice synthesis
Supabase for auth, user data, topic storage, and prompt logs
SwiftUI + RealityKit on iOS for a native, performant, and AR‑ready UI

Challenges we ran into

Managing model inference time — especially when combining text, image, and voice generation
Fine‑tuning prompts to produce outputs that are informative yet engaging
Designing a clean UI that balances educational clarity with visual appeal

Accomplishments that we're proud of

Built a functional mobile experience that generates text, visuals, and AR content for historical topics
Designed an interface that feels intuitive and immersive
Achieved high‑quality AI‑generated outputs that are contextually grounded and visually consistent
Received encouraging feedback that the experience makes history “feel alive again”

What we learned

We learned how to develop with Swift and Apple frameworks to build native, performant interfaces. We also gained experience coordinating multiple AI tools, handling database integration with Supabase, and running inference workflows end‑to‑end. Most importantly, we learned how to make advanced AI technologies accessible and meaningful through creative education.

What's next for HistorioVision

Expand support for larger local and cloud‑hosted LLMs to improve narrative depth
Add collaborative timelines where multiple users can explore and build together
Integrate richer voice synthesis for narrated historical stories
Partner with educational institutions to use this as a digital learning aid