MOMO · on-device dictation
Speak. It becomes text.
Tap the mic, talk, tap again — Whisper runs inside this browser and types out what you said.
100% on-device · nothing is uploaded
00:00
Loading…
1 · sound → picture
Your voice is resampled to 16 kHz and turned into a mel spectrogram — an image of which frequencies happen when.
2 · encoder listens
Whisper's Transformer encoder reads that image and builds a rich representation of the audio.
3 · decoder writes
The decoder generates the text one token at a time, attending back to the audio — right here on your device.