Create two PNG images of your character: one neutral and one mouth-open/expressive
Make sure both images have transparent backgrounds
Use the same canvas size for both images
Keep the character centered in both images
Save both images as PNG files
Choose PNGTuber software or a streaming app that supports image swapping
Import the neutral and expressive PNGs into the software
Set the neutral image as the default idle state
Set the expressive image to appear when your microphone detects sound
Adjust the microphone sensitivity threshold
Test the avatar by speaking and staying silent
Resize and position the PNGTuber on your stream layout
Add hotkeys or extra expressions if the software supports them
Use a clean, high-resolution design for better stream quality
Export or save your setup before going live