A social feed made only of voice — record a take, it transcribes and files itself.
with Aashna Majmudar
Takes is a voice-first social app: no text posts, no images, just short audio takes on whatever you care about. You tap the mic, say the unfiltered version of your opinion, and post it. Each take is transcribed and tagged by topic on the way in, so you can search takes by topic and browse by category even though nothing was typed.
The friends feed plays takes from people you follow with a synced transcript that highlights along with the audio. Threads let you react to what your friends said — still in voice, nested under their take.
The longer-term goal is hands-free use: manage the whole app with earphone gestures only, so recording, listening, and reacting never need the phone screen.
Built in React Native on Expo, with Firebase for auth and data, Supabase for the audio files, and a Whisper model doing the speech-to-text.