I shipped an iOS transcription app with no backend at all

Launched today after about four months. It records a conversation, tells the speakers apart, transcribes it and titles it, all on the phone. There is no server, no account and no upload, which also meant no infrastructure to build or pay for. The constraint turned out to be the product. The people who most need transcription, lawyers, journalists, researchers, therapists, are the ones who cannot put a privileged conversation into a cloud tool. That is a real market that the incumbents structurally cannot serve. Built in Swift on Apple's on-device frameworks. No servers also means no running costs, which changes what the business has to look like. apps.apple.com/app/apple-store/id6808023173 Happy to talk about anything, including the parts that went badly.