Get Multiple Consistent Characters in Midjourney Easily

A Midjourney generated image using Midjourney Automation Suite

Creating consistent characters in Midjourney can be tricky. Getting *multiple* characters to appear consistently across different images seemed hard before. This guide shows you an easy way to do it without complex methods.

Starting with the Basics

This method focuses on using the Vary Region feature. You start with one image showing the characters and modify parts of it. This helps Midjourney keep the characters looking similar.

Here's a simple prompt to begin. Ask for your characters in different poses on a plain background. This makes editing easier later.

Example Starter Prompt:

three different poses of a couple Henry and Maddie isolated on a blank white background aspect ratio 3:1 style raw

Generate an image using this. You'll pick one of the resulting images to work from.

Setting Up Your Environment

If you use Midjourney on Discord, turn on Remix mode. Type `/settings` and click the Remix button. If you're using the website, you're ready to go.

Modifying the Scene

Once you have your starting image, select "Vary Region." Use the rectangle tool to draw over the areas you want to change, like the background and perhaps the poses of two of the figures if you have multiple on a canvas.

Adjust your prompt to describe the new scene. For example:

Henry and Maddie walking through a park in Spring

Submit this change. You should see an image with your characters placed into the new setting.

If you want to explore different scenes or poses using the same character style more efficiently, a tool could help manage your variations. Consider the Midjourney Automation Suite from TitanXT to streamline your creative process.

Cleaning Up and Resizing

Often, the new image will still have parts of the old scene or the dimensions will be off. To fix this, use "Vary Region" again. Paint over the old background area you want to replace. Change the prompt back to describe a plain background.

To correct the image size or aspect ratio, use "Custom Zoom" (Discord) or "Change AI" (website). Adjust the aspect ratio, for example, to 16:9.

Changing Details Like Clothes

You can even change character clothing. Use "Vary Region" and select the specific area, like a shirt, using the lasso tool. Then, add a description of the new clothing to your prompt.

Example:

Henry is wearing a blue Woolen Jersey

Submit this, and Midjourney will update the selected area while keeping the character consistent.

Dealing with Common Issues

This method works well, but you might hit some snags. Here's how to handle them.

Characters Don't Appear After Prompt Change

Sometimes, after changing the prompt using Vary Region, your characters disappear. This happens if Midjourney ignores the original image structure too much.

To fix this, use your starting image as a reference. Get the image URL. On the website, go back to the starter image, click the menu (three lines), then "Copy Image URL." On Discord, open the image in a browser and copy its URL.

Add this image URL to the very beginning of your modified prompt, before any text. This tells Midjourney to use the image as a guide.

Example Prompt with Image URL:

[image URL] Henry and Maddie walking through a park in Spring

This should make your characters show up in the new scene.

[H3]Characters Don't Change Pose or Clothes[/H2]

If your characters appear but their pose or clothes don't change according to your new prompt, the original image might be influencing Midjourney too much. There are two ways to address this:

[LI]Adjust Image Weight: When you add the image URL, you can also add an image weight parameter at the end of the prompt. In Midjourney V6, the default is 1, and the range is 0 to 3. Try a lower value like `.5`. This gives more "weight" to your text prompt than the image reference.[/LI]

[LI]Use a Different Reference Image: If adjusting image weight doesn't work, find a different reference image entirely. Search online or in Midjourney for an image that has a pose or style closer to what you want for the scene. Use the URL of this *new* image at the start of your prompt instead of your original character image.[/LI]

Refining these images often means generating many variations and tests. The Midjourney Automation Suite from TitanXT can significantly speed up experimenting with different prompts, weights, and reference images.

Limitations of the Method

This technique is very effective for creating consistent *couples* or pairs of characters who appear together in a relationship. Most examples using this method show success with two closely related individuals.

However, it doesn't work as reliably for other groups, like a parent and child or larger unrelated groups. Keeping consistency can be harder in those scenarios. Keep experimenting to see what works for your specific needs.

Mastering character consistency takes practice and iteration. Automating parts of your workflow can give you more time to experiment and refine your images. Learn more about how the Midjourney Automation Suite by TitanXT can help.

Conclusion

Using Vary Region and strategically applying image references and weights is a powerful way to get consistent multiple characters in your Midjourney creations, especially for pairs. While there are limitations, understanding these techniques lets you push the boundaries of what's possible.

Have fun creating and keep experimenting!