The toolbar

Every control on the HUD, left to right, and what each one does.

Everything the overlay does is on one strip along the bottom. This is the whole surface — there are no menus and no hidden panels.

The session bar: End with a running timer, then Sys, Vision, Answer, Settings, Attach, Layer, Move, Click-through, Hide, and quit.
The bar, left to right. Everything the overlay does is here.

The session controls

ControlWhat it does
End · 0:00Ends the session, and shows how long it has been running. The session starts by itself when you press Start session on the setup window — there is no separate start button here. See Your first session.
SysThe system-audio channel: the interviewer's voice. This is the one to check first if nothing is being transcribed.
VisionPicks the window to watch for on-screen questions. Its tooltip shows what is being watched, when the last frame was sent, and how many were skipped.
AnswerAsks for an answer to what was just said, without waiting for the app to decide a question was asked. Use it when the interviewer phrased something as a statement.

There is one audio channel, not two. Only the interviewer is captured — your microphone is never opened, so nothing you say is transcribed or sent anywhere. See What leaves your machine.

The window controls

ControlWhat it does
SettingsOpens Settings.
AttachDetaches a panel into its own free-floating window, and docks it back again. See Detaching the panels.
LayerReaches the window behind the overlay — use it when the panel is sitting on top of something you need to click or read.
MoveRepositions the panels.
Click-throughPasses clicks straight to whatever is underneath. See Clicking through the HUD.
HideHides the overlay. It keeps running — this is not quit.
✕Quits the app.

The panels

The transcript sits in one panel and the answer in another. Each panel carries its own controls on its right edge — move, minimise, close, and a resize handle in the corner — so you can put the transcript beside the interviewer's window and the answer under your webcam, rather than accepting one fixed arrangement.