Farm Your Aura

Inspiration

Aura turns the very online question “who has the most aura?” into a playful iPhone game. The goal is to make each reading feel personal and shareable while keeping it affectionate, never cruel.

What it does

Enter a public person’s name or share a link to a public profile. Aura returns a score out of 10,000, four playful stats, a tier, a verdict, and modifiers. You can farm suggested people next, collect achievements, compare the community leaderboard, and share a reading as a styled card. The rarer someone is to farm, the more points their first readings can earn.

How it was built

Aura is a native SwiftUI app. A small Cloudflare Worker looks up factual public context; Apple’s on-device Foundation Models writes the playful reading, so the generation prompt stays on the iPhone. The score combines the model’s reading with a public footprint signal so scores have a wider, rarer range. SwiftData stores local progress, Game Center powers achievements and the leaderboard, and RevenueCat serves three optional tips.

Challenges and what I learned

The model’s four ratings tended to bunch together, which made people’s scores feel too similar. I added a public-footprint signal and tuned the scoring blend to give the distribution more range while keeping each result grounded in facts. I also kept lookup and generation separate: the server returns a short factual summary, while the on-device model is instructed not to add biographical claims. Profile search uses App Attest and avoids sensitive topics because a public profile may belong to a private individual.