AightBet
Turning group-chat trash talk into real bets you can settle.
Bets between friends happen constantly: "$20 says you won't do it". They die the same day. No one tracks them, no one settles them, and the only record is a group chat someone will scroll past tomorrow. AightBet turns that into a light structured layer: create the bet, track the outcome, settle up.
I designed and built the MVP solo using Figma + Cursor — a social betting app that sits between BeReal, Venmo, and Kalshi.
Social, financial, predictive — in one tap.
Friend bets already sit at the intersection of three behaviours people do every day: social posting, sending money, and making predictions. No single app handles that combination. AightBet's job is to make the bet itself feel as casual as a group-chat reply, while quietly handling the structure underneath.
Create a bet in fewer taps than writing one.
The create-bet flow was the hardest part to get right. It had to capture stakes, participants, and a clear resolution condition without feeling like filing a form. I iterated in Figma, pushed the design into Cursor via Figma MCP, and kept tightening the flow until creating a bet felt faster than typing it out in iMessage.
One builder, full stack, live prototype.
I ran product, engineering, and design at once. Figma stayed the source of truth for system decisions; Cursor handled implementation; small features like the image uploader were vibe-coded directly. The real bottleneck wasn't ideation — AI accelerates that — it was the translation layer between tools.
Key Lessons
- AI speeds exploration, not decisions — the hard calls still need human judgment.
- Figma stays the anchor — as soon as Cursor became source of truth, decisions got noisy.
- Tool translation is the bottleneck — the gap between Figma and code ate more time than either tool alone.
Next Project
Shopify Fulfilment Config Manager Redesign