Source transcript
—Timestamped transcript segments will appear here.
Private speech tools in your browser
Speech to editable text. No audio upload to server.
Transcribe local recordings or your microphone on-device. Translate supported non-English speech to English after transcription.
1 · Input
Choose a source
Checking runtime…
Local file
Select, drop, or paste browser-decodable audio/video where the browser exposes clipboard media.
MP3, WAV, M4A/AAC, OGG/Opus, WebM, FLAC and supported video containers depend on your browser codecs.
Microphone
AudioWorklet capture is segmented by neural VAD before progressive Whisper transcription.
Permission is requested only when recording starts.
2 · Process
Session
ReadyChoose a source.
3 · Result
Transcript
Local data
Storage controls
Only application/model caches and theme preference persist. Audio, transcript, translation and speaker content are session-only.