BuzzBroker
A fake-money trading floor for influencers.
If you found them at 10k, your portfolio is the receipt.
Everyone has a creator they followed at 5k. Everyone tells the story. Nobody has receipts. BuzzBroker is the receipt. You search a creator, and if nobody else has touched them yet, that search is their IPO. The app scrapes their numbers on the spot, sets an opening price, and you buy in at the ground floor. Other people find them later, demand drives the price, you profit. Skin in the game on the discovery flex.
Mechanics borrowed from BestBroker, energy borrowed from the part of Gen Z that has been saying "told you so" about creators since they were 13. Fake money, real calls, public leaderboard so the group chat can't pretend they didn't see it.
Currently on TestFlight. Active development is on hold while I focus on other things — but the build is shippable and I'm happy to hand out invites.
Visit buzzbroker.app →
What it does
- •Search 1M+ Instagram and TikTok creators, buy and sell shares with dynamic pricing tied to follower count, engagement, and crowd momentum
- •RuneScape-style XP from level 1 to 50, achievement badges (Scout, Early Believer, Whale, and more), daily login streaks with milestone rewards
- •Real-time portfolio tracking — P&L charts (1D / 1W / 1M), asset breakdowns, public leaderboards ranked by portfolio value
- •Follow other investors, share trade screenshots with deep links, public profiles surfacing badges and trade history
- •AI categorization tags creators by niche (fashion, gaming, tech, etc.) and infers audience demographics so discovery actually works
The architecture
Expo (React Native) + NativeWind on the front end, with Reanimated handling the chart animations and trade modals. The backend is Convex — every price tick, trade, and badge unlock is a real-time subscription, so leaderboards and portfolios update without a refresh. Creator data comes in via SearchAPI.io scrapes on a 6-hour cadence; pricing recomputes every 5 minutes from follower deltas and engagement weighting.
DeepSeek does the creator categorization — niche tags, inferred demographics (gender, age bracket, language, location) — so search and discovery actually surface relevant people instead of a wall of usernames. The whole UI follows a Swiss-minimalism style guide: typography-driven hierarchy, no gradients, no rounded cards, no emoji in the chrome. Sharp edges. Honest layouts.
Why it's parked
The TestFlight build is solid. The path from there to the actual business is not. The interesting revenue isn't the app, it's the trend intelligence layer underneath: what thousands of users collectively bet on is a real signal, and brands and agencies pay for that. Getting there means a licensed data pipeline (HypeAuditor, Social Blade), legal scaffolding around what you can and can't scrape, and enough users for the wisdom-of-the-crowd data to actually mean something.
That's a full round, not a side project. Parked until I'd rather raise on traction than promises, or until someone wants to fund the next leg.