TasteTrail Jaipur

Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for TasteTrail Jaipur

Inspiration

Travel planning often returns generic lists that ignore the way people actually describe a day: a place, a few interests, a budget, a time window, and practical constraints such as walkability. I wanted to build a small, transparent planner for Jaipur that turns those inputs into culturally relevant recommendations while keeping the origin of each result visible.

What it does

TasteTrail Jaipur accepts a location, comma-separated interests, budget, time window, and constraints. It resolves free-text interests through Qloo tags, queries the Qloo Insights API, applies a restaurant category filter when the request is food-focused, and returns a readable JSON plan with recommendation names, entity IDs, reasons, filters, and provenance.

The demo shows two live scenarios: a heritage and vegetarian-food day plan, followed by a low-budget, walkable street-food plan. The app keeps user constraints separate from Qloo-generated recommendations so the result remains explainable.

How I built it

The application is a lightweight Python web server with a dependency-free browser interface. A small Qloo client handles tag lookup and Insights API requests. The planning layer normalizes inputs, resolves tags, applies category filters, and preserves provenance. A presentation tie-break improves the ordering of clearly food-oriented venue names without changing the underlying Qloo response or pretending that the local ordering is a Qloo score.

Challenges I ran into

The main challenge was translating natural-language interests into stable Qloo tag IDs and avoiding results that matched a theme but were poor practical venues. I addressed this by resolving tags first, using explicit venue filters for food requests, and documenting the local presentation rule separately from the external recommendation source. I also kept credentials in the local environment and out of the repository.

Accomplishments

I built a working end-to-end flow from user constraints to live Qloo recommendations, added transparent provenance, handled empty-result cases, and tested the tag and filter paths with local checks. The same interface can support other cities and interest combinations without hard-coded Jaipur recommendations.

What I learned

Good recommendation UX needs both cultural relevance and explicit constraints. Showing resolved tags, applied filters, and provenance makes it easier to understand what came from the API and what was supplied or formatted by the application.

What's next

Future work includes distance-aware walking routes, time-aware sequencing, richer venue details, saved itineraries, and additional validation for constraint satisfaction across more cities.

Built With

• css
• html
• javascript
• python
• qloo
• qloo-insights-api
• recommendation-systems
• rest-api
• travel-tech