Speech to text

Turn the microphone on and speak. Text appears as you talk; a short pause ends a line.

Microphone off

Level — the line is where the server starts counting it as speech.

Echo each line back in as over a telephone line

Each line is re-spoken by /tts in the preset voice, or by /clone from the line's own audio — about 10 s a line. The telephone line is 300–3400 Hz at 8 kHz, μ-law, as a call carries it. The microphone is muted while an echo plays.

Nothing yet.