Live Transcribe — Google's Real-Time Captioning App Developed with Gallaudet University and a Deaf Lead Researcher

Live Transcribe, released as a free beta for Android on 4 February 2019, turns nearby speech into on-screen captions. It was led by Dimitri Kanevsky, a deaf Google research scientist, with engineer Chet Gnegy, and developed in collaboration with Gallaudet University researchers (including deaf professor Mohammad Obiedat) to establish which features mattered most, such as sound-event indicators for noise, name calls and similar cues.
Reported figures: captions in 70+ languages and dialects (Wikipedia cites 80+ with varying accuracy), 500 million+ downloads as of early 2023, and the speech engine open-sourced under Apache 2.0 in August 2019.
Limits are documented: it needed an internet connection until an offline mode arrived in March 2022 for devices with at least 6GB RAM and certain Pixel phones; accuracy varies across languages; and, as HearingTracker noted, human operators are still needed for some real-time captioning situations. No peer-reviewed outcome evaluation of communication benefits was found in the sources reviewed, so evidence rests on co-design process and adoption.

Read the full analysis: https://www.hearingtracker.com/news/google-live-transcribe-app