Blend-in overlay

- Press Ctrl+Shift+Space to open the overlay and choose Blend as the visual mode at the top.
- Use the eye button or Ctrl+Shift+, to make the text as faint as you can still read.
- Keep the window narrow and just below your webcam.
That's all you need the first time. Leave everything else on its defaults.
Click Resume & notes at the top and upload your latest resume. Add the job description if you have one.
Call on this computer: Video interview. Call on a phone or speaker: Speaker mode. Tests: Online test.
You get an answer you can say right away.
Before the real thing: run Practice run under Settings → Interview prep (about 1 minute). Press ? in the app anytime to see every shortcut.
Learn these 4 first. Look the rest up when you need them.
Press when the interviewer finishes. Also for follow-ups on your code.
Correct a misheard word, or type a question directly.
Multiple-choice, true/false, fill-in and short answer.
Approach, code and complexity. Press again to revise.
These are the defaults. Change any key under Settings → Shortcuts; the app always shows your current keys. On a laptop, if F1 does nothing, try Fn+F1.
Choose by where the interviewer's voice comes from.
Zoom, Teams, Meet or any call running on this computer. CalmRound hears the call audio and your mic; headphones recommended. Prefer no key presses? Switch to Auto answer in Settings → AI answer settings.
The call plays from a phone or speaker near your computer's mic, at medium volume. Press F1. For auto answer, first record about 20 seconds under Settings → My voice.
An AI interviewer asks questions and you answer as you would for real. Press F1 when you want a suggested answer.
F5 for regular questions, F4 for coding. For a question spanning several screens, press Alt+S on each, then F5. Turn on Read test answers aloud automatically before you start if you want to listen with headphones.
Problem on screen: F4. Problem only read aloud: Ctrl+F4 (Mac: Option+C). Asked why you wrote it that way: F1. Requirements changed: press the same key again. Don't press F8 mid-question.

Pick one setup, then film 10 seconds from the second camera's angle to check.


Always film a test run. From the interviewer's angle you should see your face and hands but not the answer text. What you can read isn't what the camera can read.
If you're asked to install proctoring software that inspects your computer, use Option B and keep only the test page open on the computer.
Press Ctrl+Shift+Space to open or close it. For a normal video call, change nothing.
The main window, overlay and answers are designed to stay out of screen shares, recordings and screenshots on Windows and macOS. After a system update, run Settings → Capture self-check again.
Clicking or dragging the overlay doesn't pull focus from your call, test page or IDE.
Overlay gone? Press Ctrl+` in case it's hidden. Can't click it? Press Ctrl+Shift+G. Wrong size or place? Restore the recommended size under Settings → General → Overlay.
Materials go into three tabs by purpose. Settings can stay on defaults the first time.
Resume, job description, project details. The AI builds each answer from them.
Ready-made Q&A such as your intro or why you're leaving. Used close to word for word when asked.
A full talk or presentation. Ctrl+Shift+7 brings it up to read from; no AI involved.
Set it under Settings → General. One language the whole way: pick English, Chinese or Japanese. The interviewer switches between Chinese and English (or Japanese): pick the mixed option and you also get a translation. Start a new session after changing it. Code language for F4 and F5 is set separately in AI answer settings.
Add or paste your text under Resume & notes → Script and save. During the interview press Ctrl+Shift+7 (Mac: Cmd+Shift+7) to show or hide it, Ctrl+Shift+8 / 9 to turn pages, A- / A+ for text size. It reopens where you left off and uses no minutes.
PDF, DOCX, TXT and MD are supported; run text recognition on scanned PDFs first. If company, project or tech names keep getting misheard, add them under Settings → Role & keywords; they apply from your next session.
Search a symptom or key, such as "audio", "F1" or "overlay".
On a laptop try Fn+F1. Still nothing: switch to a key combination under Settings → Shortcuts.
Check the mode: call on this computer is Video interview, call on a phone or speaker is Speaker mode. On a Mac, allow Microphone, Screen Recording and System Audio Recording under System Settings → Privacy & Security, then quit and reopen CalmRound.
Misheard words: press F3, fix them, and the answer is regenerated. Question not picked up: repeat it aloud and press F2. Whole new topic: press F8.
Press Ctrl+F1 (Mac: Option+W) to try another. The same question is answered a different way and the original stays in history.
Too early: raise the wait to 2–3 seconds. Not at all: check that AI answer settings is on Auto answer; Speaker mode needs about 20 seconds of My voice first; tests don't use auto answer.
F3 works once a video, speaker or mock interview has started. It's not available during a practice run or an online test.
Press F1 and the answer explains the code you just got: approach, complexity, trade-offs. New requirements: press F4 again (Ctrl+F4 for a spoken problem). Don't press F8 in between.
Press Ctrl+` in case it's hidden, Ctrl+Shift+G to make it clickable, and restore the recommended size under Settings → General → Overlay.
The default secure connection doesn't need shared Wi-Fi. If it still fails, choose Local network backup, put both devices on the same Wi-Fi and scan again.
That's expected in online tests: answers move to the phone. Press Ctrl+` to bring the overlay back briefly; it returns when the phone disconnects. In interviews the overlay stays; hide it with Ctrl+`.
Check that your system output is the headphones, then click Preview in the test controls. Answers are still shown as text in the overlay.
Nothing matches. Try another word.
Still stuck? Go to Settings → About & Updates and send diagnostics to support, or email [email protected] with your operating system.