Can they tell
We do not promise you will not be noticed
The short answer, before the detail
People arrive at this page wanting a yes or a no. The honest answer is that it depends on what kind of assessment you are in, and that in the cases where it matters most, the thing that gives someone away is not the software at all.
So we will not give you a promise here, in either direction. What follows is what is actually known about the question: what monitoring software genuinely does, what an ordinary video call does not do, what human interviewers report noticing, and where we stand on all of it. If you were looking for a guarantee, we do not have one, and neither does anybody selling you one.
What monitoring software really does
Proctoring is a real industry with a real technical surface, and it is worth being specific rather than vague. The mainstream tools lock the browser so other sites and keyboard shortcuts are unavailable during the session, watch the camera for gaze direction and head movement and flag prolonged looks away from the screen, listen for other devices in the room — including the wake words that trigger phone assistants — and combine everything into a per-session suspicion score with timestamps for a human to review afterwards.
In technical assessment specifically, the platforms have gone further. Keystroke cadence is analysed against a candidate’s own baseline, external tool use is tracked, multiple monitors are detected, and vendors publish accuracy figures for machine-learning models that classify code as generated rather than written. In 2026 gaze signals arrived in coding assessments too, flagging a candidate who looks away repeatedly and then resumes typing.
None of that is a browser scanning for a list of extension identifiers. It is behavioural observation, scored. And behavioural observation does not care which tool produced the behaviour.
What an ordinary video call does not do
Google Meet, Zoom and Microsoft Teams are conferencing products. They do not inspect the extensions installed in a participant’s browser, they do not score gaze, and they do not hand an interviewer a report afterwards. A meeting link is not a proctored exam and does not become one because the stakes are high.
We are saying that as a fact about those products, not as reassurance from us. It also has an edge on it: it describes the platform today, not the rules of the process you are in, not what your interviewer is looking at on their own screen, and not what any employer may add tomorrow. A statement about software is a narrow thing, and treating it as a broad one is how people get caught out.
The signal that has nothing to do with software
The flat pause
This is the one interviewers name first. Not the fact of a pause — a couple of seconds of thought is normal and always has been — but the same four or five seconds in front of every answer, whether the question was trivial or genuinely hard. Real thinking is uneven. Waiting for something is not.
Reading eyes
Eyes tracking along lines of text look different from eyes searching a memory, and on a video tile the difference is not subtle. It is the second thing on every interviewer’s list, and it does not require any tooling to notice.
The register shift
A model writes prose that sounds like documentation. When it is recited into a conversation, the gap between it and the rest of your speech is audible — tidy, complete, and unlike the way you were talking a minute earlier.
The depth cliff
The most reliable check anybody has found is not observation at all: take the polished answer and ask “why?” twice. An answer that came from somewhere else runs out at the second level, and interviewers describe that drop as unmistakable.
Notice what all four have in common. None of them is about an extension, a bot or a process running on your machine. They are about a person reciting something they did not produce — which means they are equally true of a printed sheet of notes, a phone under the desk, and a friend in a chat window. The question “can the tool be spotted” turns out to be the wrong question. The question is whether the answer is yours.
Where the answer is simply no
In a proctored exam or a locked-down assessment, this is not a fine judgement. Those sessions restrict what the browser can do, watch the camera continuously, and are explicitly built to rule out outside help of any kind. Cue is outside help.
That is not a use case we serve, and it is not one we are working towards. We build nothing aimed at proctoring or exam-monitoring software, nothing whose purpose is to keep the extension from being seen, and no arrangement in which somebody other than you answers in your place. Those are not features waiting on a roadmap. They are the ones we decided against, and asking will not change the answer.
Why we answer this page the way we do
Search this category and half the results promise that nobody will ever know. We think that promise is worth nothing, for a plain commercial reason as much as an ethical one: nobody selling it controls the interviewer, the process, or the follow-up question. A person who buys on that promise is the person most likely to be let down by it, in the one room where being let down is expensive.
So Cue is built the other way round. The window is a normal part of your browser page. It is plainly there if you share that page. We do not conceal the extension, do not disguise what it is, and tell you here, before you install it, that we promise nobody it cannot be noticed. If that makes it the wrong tool for the interview you had in mind, this is a much better place to find out than the interview.
A more useful question than this one
Most of what people actually use Cue for is not a grey area at all. Following an interviewer in a language you are weaker in, so the conversation is about your engineering rather than your listening. Catching a question that bad audio or an unfamiliar accent swallowed. Keeping a story structured instead of wandering. Typing your own question into the window mid-call, which is a person using a tool, in the open, in a conversation they are part of.
In all of those the answer stays yours and the follow-up question holds no fear, which is the real test that everything above keeps pointing at. And if you want to build the habit before it counts, run it in a mock interview with a friend first — the same window, the same cards, none of the stakes.
FAQ
The uncomfortable questions.
Can an interview copilot be spotted?+
We make no promise either way, and we would not be able to keep one. In a proctored or locked-down assessment the whole session is built to rule out outside help, and Cue is outside help. In an ordinary video call the conferencing platform does not inspect your browser — but what an interviewer notices about how you answer is a separate matter entirely, and that is where the real signal lives.
Do Zoom, Google Meet or Microsoft Teams scan for extensions?+
They are conferencing products, not monitoring products: they do not inspect the extensions in a participant’s browser, score gaze, or produce a report for the interviewer afterwards. That describes those platforms as they are, not the rules of your process and not what your interviewer is watching on their own screen.
What does proctoring software actually detect?+
It locks the browser so other sites and shortcuts are unavailable, watches the camera for gaze direction and head movement, listens for other devices and assistant wake words in the room, and combines the findings into a suspicion score with flagged moments for a human to review. In coding assessments it adds keystroke-cadence analysis, multiple-monitor detection and machine-learning models that classify code as generated.
What do interviewers say they notice?+
Four things, consistently: a uniform pause of several seconds before every answer regardless of difficulty, eye movement that looks like reading rather than thinking, prose that sounds like documentation instead of conversation, and an answer that collapses when asked “why?” twice. None of them is about software, and all of them apply equally to notes on paper or a friend in a chat window.
Do you build anything to work around monitoring software?+
No, and this is settled rather than pending. We build nothing aimed at proctoring or exam-monitoring software, nothing whose purpose is to keep the extension from being seen, and no arrangement where somebody else answers in your place. An assessment that forbids outside help is not a use case we serve.
So what is Cue for?+
Calls you are part of, where the rules allow it and the answer is still yours: following an interviewer in a language you are weaker in, catching a question that bad audio swallowed, keeping a long story structured, or asking the window your own question mid-call. It suggests a short prompt to speak from, not a script to recite — and a prompt you can defend two follow-ups deep is the only kind that survives an interview anyway.
A tool you can describe out loud, in a call you are part of.
Trippi Cue is in the Chrome Web Store. One click on the icon when the call begins, and nothing before that.