Echo turns speech into kinetic terminal text. Speak and your words type themselves out over a live waveform of the microphone signal.
The words come from the browser's built-in SpeechRecognition, with interim results so text appears as you talk. The waveform is drawn from the audio stream itself. Recognition is set to US English and is handled by the browser, so it depends on which browser you use.
It works best in Chrome. Firefox and Safari lack speech recognition, and there you get the waveform without the text.