Inspiration
Have you ever asked a chatbot for date ideas and gotten the same five tourist traps everyone goes to? We realized standard chatbots just don't understand personal taste or culture. We wanted to build something that actually gets vibes. If two people bond over a niche indie movie, their date shouldn't be at a mainstream chain restaurant. VibeMatch was born to fix this disconnect by matching you with places based on real cultural data, not just popularity.
What it does
VibeMatch is a smart date planner. You type in a movie you and your partner love, along with the city you're going out in. Behind the scenes, it figures out the exact cultural "taste profile" of that movie—meaning it looks at what kind of food, nightlife, and activities fans of that movie usually enjoy. It then plans a highly specific date night (one restaurant and one activity) that perfectly matches that exact aesthetic and mood.
How we built it
We kept the frontend simple and fast using Streamlit in Python so we could get it hosted quickly. The core engine relies on two main pieces. First, we use the Qloo API to search for the user's movie and pull its correlation data for dining and nightlife. We take that raw cultural data and pass it into Google's Gemini. Gemini acts as the date planner, using the Qloo taste profile as strict instructions to find and recommend the actual spots in the user's city.
Challenges we ran into
Connecting the Qloo API results to Gemini smoothly took some trial and error. Sometimes the cultural correlations were so specific that the system struggled to find a real-world match if the user picked a smaller city. We had to spend a good chunk of our time tweaking the instructions we sent to Gemini so it knew how to adapt a specific "vibe" into practical places that actually exist. Handling API limits and making sure the app didn't crash right before the deadline was also a race against the clock.
Accomplishments that we're proud of
We are incredibly proud that we took this from a blank screen to a fully hosted web app in just a few hours. Seeing the first successful output was a huge win—the app took a dark sci-fi movie and correctly suggested a neon-lit cocktail bar and a retro arcade. It proved our point that adding real cultural data makes a massive difference compared to just asking a standard search engine.
What we learned
We learned that context is everything. A system is only as smart as the data you feed it, and tapping into a taste graph really opened our eyes to how connected different pop culture categories are. We also got a heavy crash course in rapid prototyping with Streamlit and managing multiple API calls without slowing down the user experience.
What's next for VibeMatch
We want to expand the inputs so people can match their dates based on their favorite Spotify artists, books, or even fashion brands. We also plan to integrate a map feature so couples can see the exact walking distance between the restaurant and the activity, and eventually add links to book a table directly from the dashboard.
Log in or sign up for Devpost to join the conversation.