LiveClue floats over your interview and feeds you the next line, the one you can just say out loud. Share your whole screen and the panel stays invisible to everyone watching. Only you can see it.
Drag the divider. You share your whole screen. The panel just never shows up on theirs.
Click through what the panel actually does on a call. It listens to both sides, picks the right model, already knows your background, and answers the way you want.
Tap F10 and LiveClue pulls the question the interviewer just asked, then answers it.
It floats above Zoom, Teams, Meet, or Webex. Nobody on the call can see it, not even on a shared screen.
Always-on listening picks up the interviewer and your mic, transcribed right on your device.
Tap F10 and read the one-line answer. Want more? F12 opens it up into a few sentences.
Ctrl+Shift+A. ⌘+Enter. Alt+Space. Chrome, Zoom, Teams, the OS, and anything watching your keystrokes can all see those chords. The day some "interview integrity" feature starts flagging odd keypresses, and it will, every one of those tools gets caught. LiveClue doesn't.
Caught at the system input layer — before any app sees it.⌘F10toggle always-on listening⌘⇧↑↓scroll the chat⌘⇧Escinterrupt⌘⇧Hcompact mode⌘⌥ ←↑↓→move the panel around⌘⇧\toggle click-throughA low-level keyboard hook fires before the browser does.Ctrl+F10toggle always-on listeningCtrl+⇧+↑↓scroll the chatCtrl+⇧+EscinterruptCtrl+⇧+Hcompact modeCtrl+Alt+Arrowsmove the panel aroundCtrl+⇧+\toggle click-throughSame interview. Same keypress. Two very different DevTools consoles.
Open Chrome DevTools → Console → run addEventListener('keydown', e => console.log(e.code)) → press F10 with LiveClue running. Nothing prints. Turn LiveClue off and try again — you'll see AudioVolumeMute fire.
Your browser sees what we let it see. Which is nothing.
Your overlay never appears in screen recordings or shared screens. The interviewer sees an empty desktop.
One line under 12 words you can speak from, then bullets with the key terms already in bold.
Speech-to-text that runs on your device, with about 130 tech terms built in. It listens to the interviewer and you.
Press F9 to capture the screen, then ask. Queue up to 5 shots for a single question.
You trigger it with F9, F10, and F12. The system catches those before the browser can, so they stay invisible.
Teach it Kubernetes, Pinecone, or your internal tool names so the transcripts catch the jargon.
Run on Anthropic (Claude) or OpenAI (GPT-5). It auto-routes a fast, balanced, or deep tier by question — or you lock one in yourself.
Switch between Concise, Code, and Explanatory with one tap. Each is a full prompt you can rewrite, so every glance comes back exactly how you want it.
Pick an accent color and window background, set the opacity, and keep separate chat sessions for each role you’re prepping. It looks and works the way you like.
Drop in your resume and the target role; answers reference your real stack and experience.
The panel looks and behaves the same on both. Only the window chrome around it changes.
Your eyes hold a fixation for about 250 milliseconds. A glance answer is built to fit inside one, so you read it without breaking the conversation. A textbook paragraph makes your eyes scan, and scanning reads as "looking something up."
So LiveClue leads with a single speakable line, then backs it with bullets you only need if they dig deeper.
"Dependency injection is a design pattern in which an object receives other objects that it depends on, called dependencies, rather than creating them internally, typically facilitated by an inversion-of-control container that manages…"
Pass dependencies in instead of newing them inside.
Landed a senior role at a FAANG after four months of using LiveClue in practice loops and the real interviews. The glance line kept me talking instead of freezing.
English isn't my first language, and live technical talk used to leave me half a sentence behind. LiveClue gave me the vocabulary fast enough to keep up.
Blanked on consistent hashing mid system-design. Tapped F10, read one line, and walked the rest myself. Got the offer.
Practiced with it 50+ times before the real loop. By interview day I barely glanced at it — it had basically trained me.
I'm a PM moving into a more technical role. The glance answers helped me not look out of my depth when the questions got into the weeds.
Got an offer from a Series B AI startup at about 30% above my old comp. Worth far more than $129.
No "most popular" badge. Pick the one that fits how you'll use it.
Pay once. Yours forever.
Buy Lifetime — $129Billed annually — save 35% vs monthly
Start SubscriptionInstall in two minutes. Paste your license key. Open it before the call.