InvisAI

Caution

Disclaimer: The creator is not responsible for any consequences or actions taken by users.

Built as a free and open source version of Cluely AI.

A Windows application for secretly plugging anything into AI. The app is hidden from all screen capture using Windows APIs. This app also is invisible to focus change detectors!

NO API KEYS REQUIRED!

Features

• 🔒 Screen Capture Protection - Hidden from screen capture software
• 💻 Focus Change Prevention - Doesn't show up in any focus change detectors
• 👩‍💻 Window Cloaking - Disguise your window as Notepad
• 🎯 Area Selection - Click two corners to select any screen area
• 🌐 Modern UI - Beautiful HTML/CSS/JS interface
• 🤖 Multiple AI Checking - Double check with so many different AI models!
  • ChatGPT
  • Microsoft Copilot
  • *Gemini
  • **Qwen
  • **DeepSeek Starred Models use OCR for text detection instead of image recognition!
• ⌨ Hotkey Support - Activate the screen capture through a simple hotkey!
• 📷 Desk Scan Mode - Temporarily hide your window for any desk scan!
• 📰 AI Consensus - Get a preview of what all AIs say in one single glance!

Screenshots

Note

These screenshots taken with by temporarily disabling the hidden feature; in reality it is impossible to screenshot the app.

Copilot Auth Setup

Other Warnings

How to Use

1. Launch InvisAI.exe
   • If Windows tells you "this app has been blocked", then press "More" and "Run Anyway"
2. Complete the warning notices.
   • To work, Gemini needs to be signed in on one of your browsers, prefferebly Microsoft Edge. Your browser must also be closed when running the first AI Capture.
   • Microsoft Copilot requires you to have a .har file saved to har_and_cookies. You can does this manually, or follow the instructions on the app to sign-in and automatically grab the .har file.
   • To use Qwen and DeepSeek, Tesseract must be installed.
3. Press Start Capture or press S
4. Left click on both your corners of the question.
5. Wait for the AIs to complete and give you an answer!

Important

Note: Qwen AI and DeepSeek AI don't have a direct image AI function, so InvisAI reads the text in the image and plugs in that. Because of this, these AIs don't support image/stimulus based questions.

Known Bugs:

• No error message if Chrome is not updated
• Sometimes Copilot gives an invalid access token even after just signing in
• Gemini not working even when signed in on Edge
• Window takes time to initialize and do affinity

Building from source

1. Make sure you have Python and Git installed
2. Clone this repo with git clone https://github.com/ruskcoder/InvisAI (or download as .zip)
3. CD into your directory: cd ../../InvisAI
4. Create a virtual environment with python -m venv venv
5. Activate the virtual environment:
   • On PowerShell, run: venv/Scripts/activate.ps1
   • On Command Prompt, run venv\Scripts\activate.bat
6. Install requirements: pip install -r requirements.txt
7. Build! (make sure venv is activated)

• On PowerShell: .\build.ps1
• On Command Prompt: build.bat

8. Find your built files in /dist/InvisAI.exe