Live Transcribe and dictation, fully offline
Turn your microphone or whatever's playing through your speakers into text as it happens, using on-device Whisper AI in 99 languages. Your voice stays on the machine — nothing is streamed to a cloud.
How to use it
- Open Live Transcribe and choose your source: microphone, system audio, or both.
- Tap Space or F9 to start; speak, or let a meeting or video play.
- Tap Space or F9 again to stop — text from each take keeps accumulating.
- Save it as TXT, SRT or VTT, send it to the editor, or right-click to translate.
Mic, system audio, or both
Point it at your microphone to dictate a note, or at your system audio to capture a meeting, a call, a podcast or a video that's already playing. Turn on both at once and you get your side of a conversation and the other person's in the same running transcript.
Whisper runs on your device, so recognition keeps working on a plane, in a locked-down office, or anywhere the wifi is bad. There's no account and no monthly limit on how long you talk.
Dictation that behaves like a notebook
Space or F9 toggles listening on and off, and each take appends to the last instead of wiping it. Think out loud, stop, gather your thoughts, hit Space and keep going — the text just grows. It's the flow you want for drafting an email or capturing ideas without touching the mouse.
When you're done you can save the audio alongside the words as MP3, FLAC or WAV, so you always have the recording to check against a tricky passage.
Translate and export where you need it
Right-click the transcript to translate it into any of 16 languages, again entirely on-device, which is handy for turning a foreign-language meeting into notes you can read.
- Save as plain
TXT, or asSRT/VTTsubtitles with timestamps. - Send straight to the built-in editor to clean it up.
- Keep the recording in
MP3,FLACorWAVnext to the text.
Questions and answers
Is my voice sent to a server for transcription?
No. Live Transcribe runs Whisper AI directly on your computer, so your microphone and system audio are never uploaded, unlike cloud dictation services that stream your voice to their servers.
Can it transcribe a meeting or a video, not just my mic?
Yes. Set the source to system audio and it captures whatever is playing through your speakers — calls, meetings, videos — or pick both to get your mic and the system audio together.
What languages does it handle?
Whisper recognises speech in 99 languages, and you can right-click the finished transcript to translate it into any of 16 languages, all on-device.
Can I keep the audio recording too?
Yes. There's an option to save the recording alongside the transcript as MP3, FLAC or WAV, so you can revisit anything the text got slightly wrong.
Related tools
Get MiniMax Converter
Cross-platform desktop app. Linux free for non-commercial use; Windows & macOS one-time €20 license. No subscription, no telemetry, no account.