Visibility map
Where the overlay is hidden during screen sharing and where it can be seen — by client and OS version.
Vizir is an AI interview assistant that hears the conversation on two channels, transcribes speech in real time, and on a hotkey serves up a hint or a ready-made solution — on a screen only you can see: an overlay that stays out of browser screen sharing, or your phone when the whole screen is captured. Below is exactly what the app can do, and how it looks.
The panel switches from the keyboard. Each window handles its own job — from the live dialogue to a line-by-line code walkthrough. Here's how it looks and what it does.
Hears both the interviewer (audio from Zoom, Meet, or the browser) and you through the mic — as two tracks, natively, with no virtual audio drivers. Lines are split into "Interviewer / You", by time, in Russian and English. Speaker echo is filtered out.
A short answer to the interviewer's last question, in the context of the conversation. Not paragraphs of theory — what people actually say out loud, in a couple of seconds.
The problem came as text or an image: screenshot ⌥C or the selection ⌥⇧C, then ⌥↩. Returns code with line numbers, a walkthrough of the approach, and complexity analysis. A vision model reads the image — you can attach several frames.
When they ask you to clarify your own answer: select a passage in "Hint" or "Solution", press ⌥E — and a breakdown of exactly that part appears here, continuing your train of thought instead of starting over.
The problem statement, the last solution it produced, and the tail of the conversation are shared across every tab. So "why this way?" and "what's the complexity?" are answered against what it already produced, not from scratch in the abstract. Vizir only sees your code if you show it with a screenshot (⌥C) or a selection (⌥⇧C).
The same hints work outside interviews too — standups, webinars and client calls. Where that pays off, and what to tell the other participants, is in an AI assistant for work calls.
No mouse needed. Generating, tabs, screenshots, moving and resizing the window — it's all on hotkeys, so your hands never leave the keyboard during the interview.
Generate a hint or a solution from the current context.
Switch tabs: Transcript, Hint, Solution, Explanation.
Task snapshot — capture the problem statement or a diagram for the vision model (you can stack several).
Grab the selected text — the problem statement or a code snippet, for example.
Hide or show the overlay with a single key.
Move the overlay window up, left, down and right.
Make the window bigger or smaller.
Navigate line by line through the answer and scroll the content of the active tab.
During browser screen sharing and in most capture paths the overlay stays hidden. Where it can be visible (native clients on macOS 15+, local recording), hints move to the phone companion. Honest, per-scenario breakdowns are on separate pages.
Where the overlay is hidden during screen sharing and where it can be seen — by client and OS version.
Hints on a phone you don't share — end-to-end encrypted.
Download Vizir, sign up and get 3 days free. No card required.