Professional Interview Assistant with Invisible Screen Overlay
An AI-powered desktop tool that helps you excel in technical and professional interviews by providing intelligent, real-time assistance while remaining completely invisible to screen sharing and recording software.
Vysper.mp4
Completely Stealth - Invisible to Zoom, Teams, Meet, and all screen sharing tools Real-time AI Assistance - Instant help with coding problems, system design, and interview questions Professional Skills - Specialized modes for different interview types
- DSA (Data Structures & Algorithms) - Complete solutions with complexity analysis
- System Design - Architecture patterns and scalability approaches
- Programming - Multi-language coding assistance and best practices
- Behavioral - STAR method responses and professional scenarios
- Sales - Frameworks, objection handling, and closing techniques
- Negotiation - Strategic approaches and persuasion tactics
- Presentation - Structure, delivery tips, and visual design
- DevOps - Infrastructure, CI/CD, and deployment strategies
- Data Science - Analytics, ML approaches, and statistical methods
git clone <repository-url>
cd Vysper
brew install tesseract
brew install sox
npm install
npm start-
Clone and Setup (first time only):
git clone <repository-url> cd Vysper npm install
-
Create Your Build:
# For your current platform (recommended) npm run build # Or specific platforms npm run build:mac # macOS (.dmg + .zip) npm run build:win # Windows (.exe installer + portable) npm run build:linux # Linux (.AppImage + .deb) npm run build:all # All platforms
-
Find Your App: Built files appear in
dist/folder
# Basic builds
npm run build # Current platform
npm run build:mac # macOS (.dmg + .zip)
npm run build:win # Windows (.exe installer + portable)
npm run build:linux # Linux (.AppImage + .deb)
npm run build:all # All platforms
# Development & testing
npm run pack # Quick build for testing (no compression)
npm run clean # Clean dist/ folder
npm run rebuild # Clean + build current platform
npm run release # Clean + build all platformsBuilt apps will be in the dist/ folder:
- macOS:
Vysper-1.0.0.dmg(installer) orVysper-1.0.0-mac.zip(portable) - Windows:
Vysper Setup 1.0.0.exe(installer) orVysper 1.0.0.exe(portable) - Linux:
Vysper-1.0.0.AppImage(portable) orVysper_1.0.0_amd64.deb(installer)
- macOS: Double-click
.dmgfile β Drag to Applications folder - Windows: Run
.exeinstaller or double-click portable version - Linux: Make
.AppImageexecutable (chmod +x) and run, or install.debwithdpkg
Clean Build Process:
rm -rf node_modules dist
npm install
npm run build-
Azure Speech (for voice commands)
- Get free key from Azure Portal
- Add to
.env:AZURE_SPEECH_KEY=your_key
-
Google Gemini AI (for intelligent responses)
- Get API key from Google AI Studio
- Configure in app: Press
Alt+G
Create .env:
AZURE_SPEECH_KEY=your_azure_speech_key
AZURE_SPEECH_REGION=your_region
GEMINI_API_KEY=your_gemini_api_key| Shortcut | Action |
|---|---|
Cmd + Shift + S |
Screenshot + AI Analysis |
Alt/Option + R |
Voice Recording Toggle |
Cmd + Shift + \ |
Show/Hide All Windows |
Alt + A |
Toggle Interactive Mode |
| Shortcut | Action |
|---|---|
Cmd + Shift + C |
Chat Window |
Cmd + Arrow Up/Down |
Skills Selection (only if Interactive mode is on) |
Cmd + , |
Settings |
| Shortcut | Action |
|---|---|
Cmd+Shift+\ |
Clear Session Memory |
- Enable Interaction Mode to scroll, click, or select inside windows.
- Use
Cmd+Up/Down(in Interaction Mode) to switch skills quickly. - Click thorugh screen works only when interaction mode is disabled
- In Stealth Mode, windows are invisible to screen share & mouse.
- Invisible to Screen Sharing - Completely hidden from Zoom, Teams, Meet
- Process Disguise - Appears as "Vysper" in system monitors
- Click-through Mode - Windows become transparent to mouse clicks
- No Screen Recording Detection - Undetectable by recording software
- Screenshot OCR - Extract and analyze text from any screen content
- Voice Commands - Speak questions and get instant AI responses
- Context-Aware - Remembers conversation history for better responses
- Multi-Format Output - Clean text and code blocks with syntax highlighting
- Problem Recognition - Automatically detects interview question types
- Step-by-Step Solutions - Detailed explanations with best practices
- Code Examples - Multi-language implementations with optimizations
- Position Windows: Place Vysper windows in screen corners before sharing
- Use Voice Mode: Whisper questions during "thinking time"
- Screenshot Problems: Capture coding challenges for instant solutions
- Check Solutions: Verify your approach with AI before implementing
- Capture Requirements: Screenshot or voice record the problem statement
- Get Frameworks: Ask for architectural patterns and trade-offs
- Verify Scalability: Double-check your design decisions
- STAR Method: Get structured response frameworks
- Industry Examples: Request relevant scenarios for your field
- Follow-up Prep: Prepare for common follow-up questions
- Node.js 16+
- Tesseract OCR (
brew install tesseract) - Audio Tool (
brew install sox) - Azure Speech Services (Free tier available)
- Google Gemini API (Free quota included)
The app remembers your interview context across multiple questions:
Contribute to make Vysper the ultimate interview companion, not a cheating tool!
- New Interview Skills - Add specialized domains (Finance, Marketing, etc.)
- Language Support - Expand beyond English for global users
- Platform Extensions - Windows and Linux compatibility
- LLM Improvements - Multiple LLM Model selections for the response
- UI/UX Improvements - Enhanced interface and user experience
- Fork the repository
- Star the project if you find it useful
- Report issues for bugs or feature requests
- Submit pull requests for improvements
- Improve documentation and add examples
- Share your success stories
β Star this repo if Vysper helped you ace your interviews or you vibed with it!
