Product

A full chatbot,
hiding in plain sight.

InvisiQ is a desktop overlay that sits on top of every application on your screen, completely invisible to screen capture, screen sharing, and recording software. Capture your screen, ask questions, and get real-time streaming answers, all through a sleek interface you drive entirely with keyboard shortcuts. Bring your own OpenAI, Anthropic, or Google API key. Your screen content never leaves your device.

In motion

On screen for you. Gone for them.

invisiq.xyz · how-it-works.mp4
Workflow

Shortcut. Answer. Back to work.

STEP 01 · summon

Hit one shortcut

VS Code, Zoom, your browser, anywhere. InvisiQ drops in above whatever you’re doing, already hidden from every capture stream that’s running.

CtrlG
STEP 02 · ask

Ask it anything

Drag-select part of your screen, paste a block of code, or just type. Pick the model right from the prompt, and swap providers mid-thought.

CtrlR region capture
STEP 03 · act

Use it on the spot

Copy the snippet, paste it straight into the app underneath, or leave InvisiQ open in click-through mode and keep right on working.

CtrlV paste to app
01 · Invisible by design

They see your screen. Not your secret weapon.

The overlay renders on your physical monitor but is dropped from every capture stream: Zoom, Teams, Meet, Discord, OBS, Snipping Tool, Print Screen, and the rest. We don’t fight the OS. We use its own privacy API.

  • SetWindowDisplayAffinity(WDA_EXCLUDEFROMCAPTURE) at window creation
  • Flag re-applied on focus changes so capture exclusion stays stable
  • Works the same across Zoom, Teams, Meet, OBS, and Snipping Tool
  • Documented Microsoft API, available since Windows 10 v2004
zoom · sales sync
01import { defineConfig } from 'vite'
02import react from '@vitejs/plugin-react'
03 
04export default defineConfig({
05  plugins: [react()],
06  server: { port: 3000 },
07})
08 
09// pin the modulePreload polyfill
InvisiQ · only visible to you
Use a separate defineConfig for prod. Pin the modulePreload polyfill and bump build.target to es2022.
build: { target: 'es2022' }
02 · Multi-model

Every model. One window. Zero tab-switching.

One window, every provider. Hand a long-context recap to Gemini 2.5 Pro, ask Claude Sonnet 4 to refactor a function, drop a screenshot in front of GPT-4o. All without changing apps or losing the thread.

  • OpenAI: GPT-4o, GPT-4o Mini, o3-mini
  • Anthropic: Claude Sonnet 4, Claude Haiku 4.5
  • Google: Gemini 2.0 Flash, Gemini 2.5 Pro
  • Seven cloud models, switchable mid-conversation
OpenAI
GPT-4o
context128K
visionyes
Default
Anthropic
Claude Sonnet 4
context200K
visionyes
Best coding
Google
Gemini 2.5 Pro
context1M
visionyes
Huge context
Google
Gemini 2.0 Flash
context1M
visionyes
Fastest
03 · Screen vision

If you can see it, you can ask about it.

Hit one shortcut, drag-select any pixels on any monitor, and a vision model gives you back a structured answer in seconds. Diagrams, error stacks, contract clauses, chart data. All of it readable, none of it visible to whoever’s watching.

  • Region or full-screen capture
  • Vision on GPT-4o, Claude, and Gemini models
  • Screenshots stay on your device, sent only to the AI provider in the prompt you choose, never stored by InvisiQ
screen.vision
region captured
InvisiQgpt-4o · vision
The selected region shows a 38% week-over-week revenue lift after the Aug 12 launch, with a momentum dip on Aug 19. Likely tied to the weekend traffic dropoff visible in the dotted overlay.
04 · Reach

One keystroke away, from every app.

A dozen global shortcuts you’ll learn in a day. Summon, capture, focus, copy, paste, switch model, panic, hide. All rebindable. None of them collide with the apps you already use.

Ctrl G Toggle overlay
Ctrl S Capture full screen
Ctrl R Capture a region
Ctrl A Focus the input
Ctrl C Copy last response
Ctrl V Paste into active app
Ctrl I Stealth typing / capture mode
Ctrl P Toggle click-through
Ctrl ] / [ Next / previous model
Ctrl Q Panic, hide everything
Esc Hide overlay
05 · Context

Reads the situation.
Answers in the right shape.

No modes to pick and no templates to fill in. Just type, and optionally attach a screenshot. InvisiQ adapts the answer to what you gave it, and a local TF-IDF memory layer recalls past sessions on-device, encrypted at rest on your machine.

  • Adapts automatically to questions, code, meetings, and more
  • Per-session memory recall, fully on-device
  • Click-through keeps the answer in sight while you work

Questions & quizzes

The answer first, then a tight explanation.

Code & errors

The approach and complexity, then a full runnable solution.

Meetings & calls

A few concise talking points and action items.

Everything else

A direct, well-structured answer. One box, every task.

Get InvisiQ

Read the answer.
Skip the awkward pause.

InvisiQ is in open beta. Use the full app free for 14 days, bring your own AI key, and tell us what to build next. No card required.