What they see
What is on their screen, and what is on yours
Start with their screen, not ours
Put yourself on the other side of the call for a moment. The interviewer has a video tile of you, a participant list, a chat panel, and whatever anybody is sharing. That is the whole surface. Everything on this page is about which of those four things Cue touches, and the answer is: the fourth one, in some configurations, and none of the other three ever.
This page exists because the honest version of that answer has a “but” in it, and a “but” tucked into a footnote is worse than no answer at all.
The three places it never appears
Where the window actually lives
Cue draws its window inside the meeting page itself, in its own shadow tree. That design has a good reason — the call’s stylesheet cannot reach into it and break the layout, and nothing Cue draws can interfere with the call’s own interface. It also has a consequence, and the consequence is the entire subject of the next section: something drawn inside a page is part of that page as far as screen capture is concerned.
So the precise statement is not “nobody can see it”. It is “it is on your screen, inside one particular page, and whatever captures that page captures it too”.
Screen sharing, all three cases
You share the meeting tab
Everyone sees the window
This is the case people get wrong. The window is drawn in that page, so sharing that page shares the window, cards and all. Close it before you share, or share something else.
You share your entire screen
Everyone sees the window
A full-screen share is a picture of your monitor, and the meeting page is on your monitor. Everything on it goes out, exactly as your notifications and your open tabs do.
You share a different tab or window
They see that tab, and only it
Chrome captures the surface you picked and nothing beside it. Share the editor, the document or the deck and the stream contains the editor, the document or the deck. This is the usual shape of a coding or a case interview.
We state this in the plainest terms we can find because the alternative is somebody discovering it live. If you are going to share anything during the interview, decide beforehand which of the three you are doing, and close the window if the answer is one of the first two. It is the same habit as closing your inbox before you present — not a trick, just not being caught out by your own screen.
Four things no sharing choice affects
Where your eyes go — reading is not the same as thinking
Eyes scanning a block of text move differently from eyes searching for a memory, and it shows on camera. Interviewers describe this as one of the first things they notice, and no software setting changes it.
The pause before you speak — especially a uniform one
A card lands a few seconds after the question ends. A couple of seconds of thought before answering is normal in any interview; the same gap in front of every answer, easy or hard, is not, and that flatness is the thing people actually pick up on.
How the answer sounds — register, not content
Read a model’s paragraph aloud and it sounds like documentation rather than conversation. The shift between your own speech and a recited one is audible. This is why the cards are short, and why reading them word for word is the wrong way to use them.
The second follow-up — “why?”, then “why?” again
Interviewers who probe two levels below a polished answer find out quickly whether it was yours. Nothing on your screen helps with that, and nothing we could build would. The answer has to be something you can stand behind.
What the window shows you, and why it is small
The design follows from all of the above. Three cards are on screen at a time and the fourth pushes out the oldest, because a wall of text is a wall you have to read. Each card carries the line it grew from, so you can check it answered the question that was actually asked. A strip of recognised speech runs underneath, so a mangled name or a misheard word is visible rather than mysterious. And the answer itself is a short prompt, not a script — a structure to speak from, in your words, with your examples in it.
A copilot that produced three paragraphs would be a copilot you read out, and reading out is the thing that shows. Small is not a limitation we are apologising for. It is the only version of this that works in a live conversation.
What we do not claim
That it cannot be noticed
We make no promise in either direction, and we would not be able to keep one if we made it. Anyone selling you that promise is selling something they do not control.
That the rules do not apply
Plenty of interviews and assessments say outright that outside help is not allowed. Cue is outside help. Reading the rules for your call is your job, and only you can do it.
That the answer is right
A language model writes it from the last minutes of the call and the brief you typed. It can be confidently wrong about your field, your stack and your own history. Read it before you say it.
FAQ
The uncomfortable questions.
Can the interviewer see the Trippi Cue window?+
Not through the call itself — it is not drawn on your camera feed and nothing of it is transmitted to the meeting. But it is drawn inside the meeting page on your screen, so it is fully visible to everyone if you share that tab or your whole screen. Sharing a different tab or window shows only that surface.
Does it show up in the participant list?+
No. Cue is an extension running in your browser rather than a service that joins the call, so nothing is added to the list, nothing waits in the lobby, and no “joined the meeting” notice goes out. That is a fact about the architecture, not a claim about what anyone might infer from the conversation.
I have to share my screen for a coding interview. What should I do?+
Decide beforehand which surface you are sharing. If you share the editor or the document as its own tab or window, the meeting page is not in that stream and neither is the Cue window. If you share the meeting tab or the entire screen, the window is on show, so close it first.
Does using it change my video or my audio in any way?+
No. Cue never requests your microphone or camera and cannot alter what they send. It listens to the tab’s audio — the other side of the conversation as the call plays it back to you — and draws text on your own screen after the video has already arrived.
Can they tell from how I answer?+
Possibly, and that has nothing to do with the software. Reading eye movement, a uniform pause in front of every answer, and prose that sounds like documentation rather than speech are all audible or visible on camera, and a second follow-up question finds out fast whether an answer is yours. Cards are short for exactly this reason: they are a prompt to speak from, not a script to recite.
Is any of this announced to the other participants?+
Nothing is announced, because nothing joins and nothing is recorded — there is no bot to introduce and no recording banner to raise. That is not the same as permission. Many interviews forbid outside help, and whether yours does is a question we cannot answer for you; our position is on the responsible use page.
Know what is on your screen before somebody else does.
Trippi Cue is in the Chrome Web Store. One click on the icon when the call begins, and nothing before that.