Real-time translation built for actual conversations
FlipTalk started with a personal problem: I wanted my parents to understand my English-speaking friends when everyone was talking around the same table.
Instead of making people take turns dictating sentences, FlipTalk keeps spoken language visible as the conversation happens. Nobody has to pause the room to follow along.
The problem
Traditional translation apps interrupt group conversations. By the time someone records a sentence and waits for a result, the discussion has already moved on.
How I built it
I built FlipTalk in TypeScript and brought it to iOS and Android with Capacitor. It combines AI speech transcription with DeepL translation across more than 30 languages.
The experience is designed to stay out of the way: live transcription appears in under 100 milliseconds and translations arrive in roughly one to two seconds.
Results
FlipTalk went from web prototype to published mobile apps in two weeks and received more than 130 downloads in its first week.
Most importantly, my parents could sit with my English-speaking friends and follow the conversation in real time.
The result is a complete AI product, designed and delivered from prototype to production.
