Operating modes

In short: the operating mode determines where Mockingbird gets the interviewer's questions — system audio, microphone, or both streams — and how it displays hints. Below: configuring the question source, testing, and Invisible Mode.

Question source

System audio. The app captures the other party's audio directly from the system — works for Zoom, Meet and Teams calls.
Microphone. Your own speech is recognized — handy for practising answers and self-checking.
Mixed. Your voice and the other party's voice at the same time; the app separates the streams.

Configuring the source

Open Settings in the toolbar.
In the source section, pick the capture device (system audio or microphone).
Check the signal level — the activity indicator should react to voice.

Testing

Press Start — the status changes to running.
Say a test phrase: the draft appears on the Interview tab.
Make sure the recognized text matches what you said.
Ask a question by voice — check that the answer arrives within the expected time.

If no text appears — check the selected device and signal level, and that the app isn't paused (the Mute button).

Invisible Mode

A separate privacy mode: the assistant window is excluded from screen capture and never gets into streams or recordings. Enabled in settings or with a hotkey.

Common issues

Silence in the draft. Wrong device selected or the system blocks capture — check permissions.
Too much noise text. Lower the VAD sensitivity or switch to the microphone.
Answers lag. Check that the GPU is active and reduce the STT model size.