Inspiration
Group nights out are hard to plan — everyone loves different things, and the loudest voice usually wins. We wanted an agent that plans from the whole group's taste, not one person's.
What it does
Girls' Night Out is a women-centred taste agent. Each friend names one thing she loved — a film, an artist or a place — and the agent resolves each to a real Qloo entity, finds the overlap in the group's taste graph, and plans one night: a dinner place, a film, and the music to match, with the reasoning explained.
How we built it
Plain HTML, CSS and vanilla JavaScript, no backend. The agent calls Qloo's Taste AI API — /search to resolve each pick, /v2/insights across place, movie and artist — and scores candidates by shared affinity plus a bonus for overlap, so genuine group taste wins. Taste data was pre-fetched so the demo runs without a key in the browser.
Challenges we ran into
Qloo's /search returns every kind of entity with a name — Taylor Swift the person, not the artist; a podcast called The Comedy Cellar, not the club — so seed resolution had to prefer the right entity type. Place results also needed a restaurant tag filter, or the agent returned museums and parks instead of dinner spots.
What we learned
Cultural grounding is the difference between a generic recommendation and a good one: the plan only works because Qloo knows what fans of each pick also love.
Log in or sign up for Devpost to join the conversation.