I Finally Fixed ChatGPT Image Ratio with Stable Diffusion Outpainting

One thing that kept bugging me:

ChatGPT’s image-1 model doesn’t support common aspect ratios like 16:9 or 9:16, which is kind of a problem when you’re making YouTube thumbnails or Shorts.

As the founder of AIThumbnail.so, I really wanted users to get clean, crisp 1920x1080 thumbnails without messing around with weird crops, white padding, or manual resizing.

So I built a fix.

Here’s the little workflow I came up with, it combines smart stretching + outpainting (using Stable Diffusion) to get perfect 16:9 images:

🛠️ How it works:

✅ Takes the original image from ChatGPT

✅ Stretches it (only slightly, no weird distortion)

✅ Figures out what pixels are still missing

✅ Uses Stable Diffusion to outpaint and fill the edges

✅ Outputs a clean 1920x1080 image, no loss, no borders, no fuss

Why this matters:

If you're working with AI-generated images, you’ve probably run into the same issue:

• Cropping ruins important parts of the image

• Padding looks bad or amateur

• Wrong aspect ratios kill your flow on YouTube, Shorts, etc.

The result: Now every image on AIThumbnail.so is auto-fixed and ready to use, no editing needed.

If you're working on anything similar or want to automate this kind of fix, feel free to ask,

happy to share how I set it up.

Thanks for reading!