Overview
Thingly is a mobile app I am building on my own. I started it to get deep practice with AI agent workflows on a real product surface, in a different stack. Most of the work is product judgment: what the app should do, how the UX should feel, and which platform constraints (StoreKit, subscriptions, review rules) should shape the architecture.
What I Am Building
- A native iOS app with Swift and SwiftUI as the core stack.
- Purchase flow with StoreKit, where the plumbing forces early product tradeoffs.
- A workflow where agents handle implementation while I own scope, UX, review, and merge decisions.
- Lightweight local development: no shared backend, no docker-compose, so parallel agent worktrees stay isolated (playing around with Conductor on this one)
Why It Exists
Mobile is a different constraint surface from the web tools I usually ship. StoreKit, App Store review, and on-device UX feedback loops do not forgive vague product calls. I wanted a project where I could practice steering agents through those constraints instead of only typing Swift myself.
The bet is simple: if agent-assisted product engineering is the new default, I should know where agents help on mobile and where human judgment still owns the outcome.
The betUse a real iOS product to learn where agent workflows accelerate mobile work and where product taste still has to stay human.
Status
Thingly is actively in development. There is no public landing page, TestFlight link, or App Store listing yet.
What is in progress:
- Core app flows and UX iteration.
- StoreKit integration and subscription modeling.
- Agent-assisted implementation with review before merge.
- Internal beta via TestFlight
What is intentionally not public yet:
- External beta access coming soon
- Marketing site or App Store presence until the product surface is ready to share.
- iOS
- Swift
- Product
- AI tooling
