Figma: Design a YouTube thumbnail with Weavy AI

Step-by-step

1. Generate initial characters and create variations for specific actions like running.
2. Isolate characters to work on them individually rather than in a single image.
3. Remove backgrounds from the isolated characters using the available tools.
4. Use style references to add specific details like boots or props to the characters.
5. Open the compositor node to move layers around on a canvas.
6. Adjust the environment, perspective, and camera angles.
7. Modify the color scheme to match your specific brand or podcast aesthetic.
8. Upscale the final composite and add typography-based text.

Options

• Use style references to apply specific aesthetics to characters
• Change camera angles (e.g., going lower) to alter the shot perspective

Watch out for

• This is not a one-click solution; it requires manual layering and design thinking.

Tips

• Work on characters individually instead of trying to generate both in one image to maintain better control.
• Design in layers and history track rather than just clicking 'regenerate' repeatedly.
• Avoid getting caught up in finding the 'perfect prompt' by using the compositor to design manually.

Highlights

“This is not a one-click solution. This is just me working here in my own brain... do everything in layers. Control this, move this.”