LocalScribe Keeps Transcription in the Browser

Field notes

The page presents transcription as a browser workflow: select or record media, choose a language and speed/accuracy setting, then edit the resulting text. It lists MP3, WAV, M4A, OGG, WebM, and MP4 support, with TXT, SRT, and WebVTT exports. Optional speaker labels distinguish voices, while the page says their models are downloaded separately the first time they are used. The site identifies its speech-recognition engine as open-source Whisper and says processing remains on the device.