Someone just open-sourced a Cluely alternative.

OpenCluely is a free Electron app that gives you real time AI assistance during technical interviews. No telemetry. No company watching your sessions. Full source code on GitHub.

Here’s what it does:

⇢ Invisible overlay that lets clicks pass through Your screen sharing stays clean. The interviewer sees nothing.

⇢ Gemini Vision instead of OCR It reads your screen visually rather than parsing text. Fewer failure points, more context.

⇢ Persistent session context It remembers what happened earlier in the interview. Your answers stay consistent across the whole session.

⇢ Multi-language support Works across stacks. Not locked to one language.

⇢ Zero telemetry No usage data. No third-party calls. Nothing leaves your machine.

Cluely raised millions selling “undetectable AI assistance” as a premium product.

OpenCluely does the same thing for free.

And you can read every line before you run it.

That distinction matters more than people are admitting.

GitHub: http://github.com/TechyCSR/OpenCluely…

TechyCSR/OpenCluely

Source: https://github.com/TechyCSR/OpenCluely

🌟 Why OpenCluely?

🖼️ Modern UI Features

📱 Interactive Windows

• Floating Overlay Bar: Compact command center with camera, mic, and skill selector
• Draggable Answer Window: Move and resize AI response window anywhere
• Close Button: Clean × button to close answer window when needed
• Auto-Hide Mic: Microphone button appears only when a speech provider is available
• Interactive Chat: Full conversation window with markdown support

🎨 Visual Design

• Glass Morphism: Beautiful blur effects and transparency
• Adaptive Layout: UI adjusts based on available services
• Smart Resizing: Windows resize automatically to fit content
• Professional Look: Mimics system applications for perfect stealth

🎯 Functional Overview

📋 Core Components

⚙️ Configuration

The setup script automatically handles configuration. You only need:

# Required: Google Gemini API Key (setup script will ask for this)
GEMINI_API_KEY=your_gemini_api_key_here

# Optional: Speech Recognition (pick one provider)
SPEECH_PROVIDER=whisper

# Azure option
AZURE_SPEECH_KEY=your_azure_speech_key
AZURE_SPEECH_REGION=your_region

# Local Whisper option
WHISPER_COMMAND=whisper
WHISPER_MODEL_DIR=.whisper-models
WHISPER_MODEL=base
WHISPER_LANGUAGE=en
WHISPER_SEGMENT_MS=4000

Note: Speech recognition is completely optional. If no configured provider is available, the microphone button will be automatically hidden from all interfaces.

🚀 Quick Start & Installation

⚡ Three Simple Steps (All Operating Systems)

1. Clone the repository

   git clone https://github.com/TechyCSR/OpenCluely.git
   cd OpenCluely
   

2. Get your Gemini API key (Required)

   • Visit Google AI Studio
   • Click “Create API Key”
   • Copy the key (you’ll need it in step 3)

3. Run the setup script (One command does everything!)

   ./setup.sh
   

That’s it! The setup script will:

• Install all dependencies automatically
• Create your .env file from env.example if needed
• Set up a local Whisper virtualenv in .venv-whisper
• Configure .env to use local Whisper by default
• Build the app (if needed)
• Launch OpenCluely ready to use (if not works use npm install & then npm start)

💻 Platform-Specific Notes

• Windows: Use Git Bash (comes with Git for Windows), WSL, or any bash environment
• macOS/Linux: Use your regular terminal
• All platforms: No manual npm commands needed - the setup script handles everything
• Windows Whisper path: setup.sh now writes WHISPER_COMMAND=.venv-whisper/Scripts/whisper.exe
• macOS/Linux Whisper path: setup.sh writes WHISPER_COMMAND=.venv-whisper/bin/whisper

🎛️ Setup Script Options

./setup.sh --build          # Build distributable for your OS
./setup.sh --ci             # Use npm ci instead of npm install
./setup.sh --no-run         # Setup only, don't launch the app
./setup.sh --install-system-deps  # Install sox for microphone (optional)
./setup.sh --skip-whisper  # Skip the local Whisper bootstrap

🔧 Optional: Speech Setup (For Voice Features)

Voice recognition is optional. You can use either Azure Speech or local OpenAI Whisper.

For the local Whisper path, ./setup.sh now handles the full repo-local setup:

1. Creates .venv-whisper

2. Installs openai-whisper

3. Points .env at .venv-whisper/bin/whisper

4. Creates .whisper-models

5. Runs npm run test-speech

6. For Azure Speech:

   • Visit Azure Portal
   • Create a Speech Service
   • Copy your key and region

7. For local Whisper:

   • Run ./setup.sh --install-system-deps
   • Or install required audio tools such as ffmpeg and sox yourself
   • On Windows, install audio tooling separately and prefer Git Bash or WSL for setup.sh

8. Add one provider to your .env file:

   GEMINI_API_KEY=your_gemini_api_key_here
   SPEECH_PROVIDER=azure
   AZURE_SPEECH_KEY=your_azure_speech_key
   AZURE_SPEECH_REGION=your_region
   

   GEMINI_API_KEY=your_gemini_api_key_here
   SPEECH_PROVIDER=whisper
   WHISPER_COMMAND=whisper
   WHISPER_MODEL_DIR=.whisper-models
   WHISPER_MODEL=base
   WHISPER_LANGUAGE=en
   WHISPER_SEGMENT_MS=4000
   

9. Restart the app - microphone buttons will now appear automatically

🎮 How to Use

🖱️ Main Controls

🎯 Workflow

1. Start OpenCluely → App appears as system process (Terminal/Activity Monitor)
2. Position Windows → Drag overlay and answer windows to preferred locations
3. Capture Questions → Use screenshot (⌘⇧S) or voice commands
4. Get AI Answers → Instant responses in draggable answer window
5. Interactive Chat → Type or speak for detailed conversations
6. Stay Stealth → All operations invisible to screen recording

🔧 Advanced Features

🎨 Window Management

• Draggable Interface: Click and drag any window to reposition
• Auto-resize: Windows automatically adjust to content
• Close Button: Click × to close answer window
• Always on Top: Windows stay above all applications

🧠 AI Intelligence

• Context Awareness: Remembers entire conversation
• Code Detection: Automatically formats code blocks
• Language Specific: Tailored responses for selected programming language
• Session Memory: Maintains context across multiple questions
• Image Understanding: DSA prompt is applied only for new image-based queries; chat messages don’t include the full prompt
• Multi-monitor & Area Capture: Programmatic APIs allow targeting a display and optional rectangular crop for focused analysis

🔊 Optional Voice Features (Azure Speech / Local Whisper)

• Chunked Local Transcription: Local Whisper transcribes short recorded segments on your machine
• Real-time Transcription: Azure Speech supports live interim recognition
• Listening Animation: Visual feedback during recording
• Interim Results: Available with Azure Speech
• Auto-processing: Instant AI responses to voice input ]

🙏 Acknowledgments

• Google Gemini: Powering AI intelligence

• Azure Speech / Whisper: Optional voice recognition

• Electron: Cross-platform desktop framework

• Community: Amazing contributors and feedback

• Vysper: UI and code structure inspiration — see Vysper by varun-singhh