Transcribe
Get a searchable text version of any recording, on your device.
Speech and image models that download once and then run on your own device: transcribe audio or video, translate subtitles, clean the noise out of a recording, sharpen an old photo, blur people out of a shot, or have text read aloud. The one exception is Meeting Summary's optional AI summary, which sends the transcript to a server — and only when you tap the button.
Get a searchable text version of any recording, on your device.
Reach an Arabic or English audience with the same video.
Turn a long recording into decisions and action items.
Bring back detail in a blurry or low-resolution picture.
Make a noisy voice recording easy to listen to.
Hide people and private details before you share a photo.
Listen to a long text instead of reading it.