A revolutionary ADHD evaluation tool through advanced pupillometry technology using eye tracking and Go/No-Go task performance analysis.
ZapGaze uses cutting-edge pupillometry technology to detect potential ADHD indicators through precise eye movement and pupil response analysis. Our non-invasive, scientifically-backed approach provides quick and accurate assessments that can help identify attention-related challenges early.
Access the deployed application:
- Web Application: http://20.74.82.26:5173
- Backend API: http://20.74.82.26:8000
To use the application:
- Visit the web application URL above
- Download the ZapGaze Agent directly from within the application (the app will guide you through the download process)
- The agent will automatically connect to the backend and enable eye tracking
The system consists of three main components:
- Frontend Web Application (Deployed on Azure VM at
http://20.74.82.26:5173): Main user interface for the ADHD assessment - Backend API Server (Deployed on Azure VM at
http://20.74.82.26:8000): Main database and business logic - Local Agent Server (
http://localhost:9000): Eye tracking and camera operations (runs on user's local machine)
- Webcam for eye tracking
- Windows or macOS computer
- The ZapGaze Agent will be downloaded automatically when you start the assessment
Go to http://20.74.82.26:5173 in your web browser.
- When you start the assessment, the application will prompt you to download the ZapGaze Agent
- Click "Download Agent" and select your platform (Windows or macOS)
- The agent file will download automatically
For macOS:
- Unzip the downloaded
ZapGazeAgent.zipfile - Right-click on
ZapGazeAgentand select "Open" (you may need to do this twice if macOS blocks it the first time) - Important: If macOS shows a security warning, go to:
- System Settings → Privacy & Security
- Scroll down to find "ZapGazeAgent" or look for a message about allowing the app
- Click "Allow" or "Open Anyway" to permit the app to run
- The agent will start and display: "Agent URL: http://localhost:9000"
For Windows:
- Double-click
ZapGazeAgent.exe - If Windows Defender shows a warning, click "More info" then "Run anyway"
- The agent will start and display: "Agent URL: http://localhost:9000"
- Return to the web application (http://20.74.82.26:5173)
- Complete the intake questionnaire
- Begin your ADHD evaluation
Note: Keep the agent running in the background while you complete your assessment. You can minimize the agent window, but don't close it.
-
Agent Won't Start on macOS
- Go to System Settings → Privacy & Security
- Look for "ZapGazeAgent" in the list and click "Allow"
- If you don't see it, try right-clicking the app and selecting "Open" again
- You may need to restart your computer after allowing the app
-
Camera Access Denied
- Ensure your browser allows camera access
- Check that no other application is using the camera
- On macOS: Go to System Settings → Privacy & Security → Camera and ensure your browser is allowed
-
Port Already in Use
- Another instance of the agent may be running
- Close other instances or restart your computer
- On macOS/Linux:
lsof -ti:9000 | xargs kill - On Windows: Check Task Manager for
ZapGazeAgent.exe
-
Agent Not Connecting
- Ensure the agent is running and shows "Agent URL: http://localhost:9000"
- Check your internet connection
- Try restarting the agent
The application is deployed on Azure Virtual Machine:
- Frontend: http://20.74.82.26:5173
- Backend: http://20.74.82.26:8000
- Agent: Users download and run locally (connects to backend automatically)
The deployment uses Docker Compose to run the backend and frontend services. The local agent is packaged as a standalone executable using PyInstaller and distributed via GitHub Releases.
To build a new version of the agent executable:
# macOS/Linux
cd agent
./build.sh
# Windows
cd agent
build.batThe executable will be in the dist/ directory. See agent/BUILD.md for detailed build instructions.
For support and questions, please contact joumanasalahedin@gmail.com or somtochukwumbuko@gmail.com
For issues with the agent or application, please open an issue on GitHub