Face Blur — Tutorial

🎥 Webcam Mode

Real-time face blurring for live video feeds, perfect for video calls and streaming.

1

Navigate to Webcam Mode

Click "Webcam" in the navigation or go directly to the webcam page.

2

Grant Camera Permission

Click "Start Webcam" and allow browser access to your camera when prompted.

3

Adjust Settings

Configure blur intensity, detection rate, and resolution to suit your needs.

4

Record (Optional)

Click "Start Recording" to capture the blurred output. Download when finished.

💡 Tip: For better performance in video calls, reduce output resolution to 50% and increase the detection update rate to 300–400ms.

🎬 Video Upload Mode

Process pre-recorded videos with face blurring and download the results.

1

Navigate to Video Mode

Click "Video" in the navigation menu.

2

Upload Your Video

Click the upload area or drag and drop your video file. Supports MP4, WebM, AVI, and MOV.

3

Preview & Adjust

The video will start playing automatically with blurred faces. Use playback controls to navigate.

4

Record Output

Click "Start Recording Output" to capture the processed video.

5

Download

Click "Stop & Download" when finished. Your blurred video saves as a WebM file.

⚠️ Note: Recording is limited to 120 seconds (2 minutes). For longer videos, record in segments.

⚙️ Settings Explained

Setting	Description	Recommendation
Output Resolution	Controls canvas output size. Lower = better performance.	50% for most uses, 25% for older devices
Blur Radius	Intensity of the blur effect. Higher = more blur.	40–60px for adequate privacy
Detection Update Rate	How often AI detects faces (ms). Higher = less CPU.	200ms smooth, 400ms+ for performance
Detection Sensitivity	Confidence threshold. Higher = fewer false positives.	0.75 (default) works for most scenarios

ℹ️ Performance: If you experience lag, lower resolution to 25%, increase detection rate to 400–500ms, and close other applications.

🎯 Best Practices

For Real-Time Webcam Use

Use good lighting for better face detection accuracy
Position yourself front-facing to the camera
Avoid rapid head movements for smoother blur tracking
Test settings before joining video calls
Close unnecessary browser tabs and applications

For Video Processing

Use original resolution (100%) for final outputs
Preview the entire video before recording
Adjust blur radius based on video quality and face size
For long videos, process in 2-minute segments

Privacy Considerations

All processing happens locally in your browser
No video data is uploaded to any server
Clear browser cache if working with sensitive content
Blur effect may not be perfect in all scenarios — always verify

🔧 Troubleshooting

Camera not working?

Check browser permissions for camera access
Ensure no other app is using the camera
Refresh the page and grant permission again
Try a different browser (Chrome offers best performance)

Performance issues?

Lower output resolution to 25%
Increase detection update rate to 400–500ms
Close other browser tabs and applications

Faces not detected?

Ensure adequate lighting on faces
Lower the detection sensitivity setting
Make sure faces are clearly visible and front-facing
Check that the AI model loaded (status indicator shows "Ready")

ℹ️ Still having issues? Visit the support page for more help or to report a bug.

❓ Frequently Asked Questions

Is my video data sent to a server?

No. All processing runs locally via TensorFlow.js. Your video never leaves your device.

Which browsers are supported?

Modern Chrome, Firefox, Edge, and Safari. Chrome generally offers the best performance.

Why is there a 120-second recording limit?

This prevents excessive browser memory usage. Record in multiple segments for longer videos.

What AI model is used?

Google's BlazeFace model via TensorFlow.js — fast, lightweight, and runs entirely in the browser.