WhatsApp Spotted Working on Feature to Transcribe Voice Notes on Latest Android Beta Version

WhatsApp has been spotted developing a feature that transcribes voice notes on Android devices. Using the speech recognition feature on your smartphone, WhatsApp will locally process voice notes that are sent and received, displaying the transcribed text in the same chat window. The feature, currently in development, could come in handy when it is not feasible to listen to a voice message. It might also improve accessibility for hearing impaired users, enabling them to use the built-in voice notes feature.

Feature tracker WABetaInfo discovered a new feature in development on WhatsApp beta for Android 2.24.7.8 — the ability to transcribe voice notes. The Meta-owned chat platform is still working on the feature, which means that even beta testers can't try it out at the moment. However, the feature tracker has shared an image of what the feature might look like when it is released.

WhatsApp will download additional data to enable voice note transcripts
Photo Credit: WABetaInfo

In the screenshot published by WABetaInfo, we can see a pop up card that explains how the feature will work when it is eventually rolled out to users. A message informs users that the app will need to download 150MB of new app data to their smartphone, in order to enable voice note transcripts on the smartphone.

WhatsApp will also inform users that it will use on-device speech recognition, which will allow the app to provide voice note transcripts while protecting user privacy and preserving its end-to-end encryption for messages. This also suggests that voice notes that are already downloaded can be transcribed even when no data is available.

It's worth noting that WhatsApp has been testing the same feature since May 2023, when version 23.9.0.70 of WhatsApp beta for iOS was rolled out. The app also downloads these language packs onto users' smartphones, enabling transcription in different languages. WhatsApp users on iOS can also search through these transcripts as the text is indexed like regular messages, according to the feature tracker.