Inspiration

As a non-soccer fan, it was hard to keep up with the World Cup, especially with the number of prominent players coming from soccer clubs and organizations around the world. We wanted to build something that would make it easier to follow news about the tournament in one place.

What it does

World Cup Buzz is a companion app for the tournament, built around three core features:

Match Timetable : a clean, at-a-glance view of every World Cup fixture: who's playing who, and when. Team Voting :for any given match, users pick the team they're backing, turning passive schedule-browsing into an active, personal experience. Team Comparison & News : side-by-side comparison of the two competing teams, along with the top current headlines for each, summarized in plain language so fans get context without needing to read five separate articles.

The goal was to compress "figure out what's going on with this match" from ten browser tabs down to one screen.

How we built it

We prototyped on Lovable, which let us move quickly from concept to a working interactive app without splitting our limited hackathon time between backend plumbing and product decisions.

Match and team data is pulled from a sports data API, giving us live fixture information rather than static/mock data. News aggregation and summarization uses the Claude API .we pull recent headlines per team and prompt the model to return structured output: overall tone, key storylines, and a plain-language summary a casual fan can read in seconds. We deliberately designed the prompt to avoid inventing facts not present in the source headlines, since accuracy mattered more to us than making the summaries sound punchier. We split responsibilities across the team: one teammate focused on the product framing and pitch narrative, one on walking through and refining the user journey in the prototype, and one on the technical integration between the data sources and the app itself.

We designed the underlying data model to be sport-agnostic — match, team, and vote as generic entities ,even though we scoped the actual MVP to the World Cup only, so extending to other tournaments later would be a data change rather than a rebuild.

Challenges we ran into

Streaming was a dead end. Early in the hackathon, we explored the idea of letting people watch matches directly in the app. We quickly learned this runs into serious licensing and security constraints — live sports video is tightly controlled by broadcasters — so we pivoted to a second-screen companion model instead: World Cup Buzz complements however someone is already watching the match, rather than competing with the broadcast itself. Scoping discipline. We originally wanted to support multiple tournaments (World Cup, Wimbledon, and others) simultaneously. With limited build time, we made the deliberate call to scope the MVP to the World Cup alone, and keep the data model flexible enough to extend later rather than trying to build breadth and depth at the same time. Keeping AI-generated content trustworthy. Feeding real news headlines into an LLM for summarization carries real hallucination risk, especially with sports storylines where details (injuries, transfers, controversies) matter. We had to be deliberate about prompt design — instructing the model to stay grounded in the source material and default to a neutral read rather than guessing when coverage was thin or ambiguous.

Accomplishments that we're proud of

The team news comparison feature came together better than we expected. Turning raw headlines into a clean, plain-language, side-by-side read is the kind of feature that sounds simple but is easy to get wrong — we're proud of how legible and useful the output ended up being. We made a hard scoping call and stuck to it. Cutting multi-sport support and live streaming mid-hackathon wasn't easy, but it meant we finished with a polished, focused MVP instead of a half-built one spread across too many ideas. We designed for the future without overbuilding for it. Keeping the data model sport-agnostic means World Cup Buzz isn't a dead end after this tournament — it's a foundation we could genuinely keep building on.

What we learned

How much product value comes from reducing the number of steps between "I'm curious about this match" and "I understand what's going on" — the comparison + news feature ended up being the feature we were proudest of, even though it was the simplest to explain. The value of scoping down early. Cutting multi-sport support from the MVP was a hard call to make mid-hackathon, but it let us actually finish and polish the core experience instead of shipping something broader and half-built. Prompt design matters as much as the feature idea itself — the same "team news summary" feature could easily have become inaccurate or misleading without careful constraints on the model's output.

What's next for World Cup Buzz

We'd like to extend World Cup Buzz beyond the World Cup to other major single-elimination tournaments, and explore deeper live-match engagement (real-time event tickers, in-match polls) as a way to keep the app relevant during matches themselves, not just before and after.