I built a Slack-native World Cup prediction pool for distributed teams
Slack-native World Cup pool, but office brackets already solve this.

Zero-signup tournament links beat the usual email-gate for casual pools.
Friends and family groups organizing casual sports betting pools
OfficePools · Fantasy Premier League · Bracket Challenge
I was testing Cursor for the past week and a half and I decided to build a quick platform for my family and friends to make a little prediction tournament for the World Cup 2026.
My goal was to have the easiest possible setup for elderly relatives, no sign ups, mobile friendly, $0 budget for hosting platform.
I doubt this will get popular but in the odd case of happening just letting you know I'm using free vercel and neon db tiers :D In case you wonder why is slow and/or down.
I'd love to get some feedback on the scoring system, do you think I can improve it and make it more competitive?
Tech stack & development: - vibe coded with mainly Opus 4.7 with Cursor - Neon free tier 0.5 GB db - vercel free tier - github actions - for games - I've loaded the whole tournament and found API - football-data.org cron for pulling the results (free API key for 10 request per min) enough for what we have here
- React - Next.js - translated to 4 languages English/German/Spanish and of course Bulgarian.
Thanks, Anton
Slack-native World Cup pool, but office brackets already solve this.
Clean prediction league app, but Excel sheets already do this for free.
Clean JSON for 92 years of World Cup data when official APIs are expensive.
Works fine, but ESPN and FIFA already offer calendar feeds for matches.
Solar position data for stadium seats when SeatGeek and StubHub don't track sun exposure.
The smart bit is packaging the entire tournament as callable tools — timezone-aware schedules, zero-query briefings that auto-detect the tournament phase, head-to-head histories, visa rules and city guides all come pre-bundled so agents don't need external APIs. Install with npx and there's a Try playground plus ChatGPT/Telegram wrappers, which makes experimentation trivial; the only real limitation is that the build-time data will need updates as qualifiers and logistics change.