CulturePilot

CulturePilot Discover Culture Beyond the Tourist Traps

Inspiration

Discovering a city is about more than finding popular places. Music, food, art, fashion, and local creative communities all shape how a place feels, yet conventional search tools often return generic recommendations that don't reflect someone's cultural interests.

I built CulturePilot to explore a better approach: helping people discover cultural connections based on their tastes, interests, and the creative identity of a destination.

What it does

CulturePilot is a cinematic cultural discovery experience powered by the Qloo Taste Graph API. Users describe the atmosphere and interests they want to explore, and the application turns that intent into cultural recommendations and connections.

The experience combines an interactive visual stage with cultural discovery results, affinity signals, and semantic filtering diagnostics. Instead of presenting results as guaranteed facts about a place, CulturePilot aims to make the connections behind its recommendations easier to explore.

For example, a discovery query focused on Lagos can surface cultural entities such as Terra Kulture and Crepaway Nigeria, alongside other taste-related connections returned by Qloo.

How i built it

I built CulturePilot as a web application with an interactive frontend and a server-side API integration for Qloo. The interface combines cinematic video elements, interactive controls, and a results drawer to make cultural exploration feel engaging rather than like a conventional search form.

The application processes user intent, queries the cultural graph, and presents returned entities with affinity information and filtering diagnostics. We also added explicit error handling, request timeouts, retry behavior, and transparent empty states so that unsuccessful or degraded queries do not become fabricated recommendations.

Challenges I faced

One important challenge was destination extraction. Certain alphanumeric destination inputs were not captured correctly, which could trigger an unscoped query and a long wait. We improved the extraction logic and added regression tests for standard cities, multi-word destinations, and alphanumeric inputs.

I also worked on making failure states useful. When the cultural graph cannot provide matches, CulturePilot now communicates that limitation instead of filling the interface with empty recommendation sections or inventing results.

Finally, i needed to balance a cinematic, interactive experience with reliable API behavior and clear explanations of what the data does and does not prove.

What I learned

I learned that an engaging AI-powered discovery product needs more than an attractive interface. Reliable intent handling, grounded data, transparent failure states, and clear communication about recommendation signals are essential to user trust.

CulturePilot is our exploration of how cultural knowledge graphs can make discovery feel more personal, connected, and meaningful.

Try CulturePilot

Live demo: https://culturepilot-one.vercel.app Source code: https://github.com/JimmyOgb/CulturePilot

CulturePilot presents cultural affinity as a discovery signal, not a guarantee of venue availability, current events, or personal preference.