
Create Children’s Books with Consistent Characters Using Midjourney, ChatGPT, and Canva
May 1
5 min read
0
78
0

Making children’s books can be quite an effort. But now, tools powered by AI like Midjourney make the process easier and more fun!
This guide shows you step-by-step how you can create a children’s storybook with the same character appearing throughout. We will use features like Midjourney’s Character Reference, along with ChatGPT for the story and Canva for the layout.
Let’s see how AI is changing things for authors who want to create books for children.
Story Ideas & Script with ChatGPT
We start by using ChatGPT to get ideas for children's stories. You can ask it questions like, "What common themes do children's books often have?"
Once you get some ideas, pick a theme you like. For example, self-confidence and self-expression. Then, ask ChatGPT to write a short story. Tell it the age range for the kids (like 4 to 6) and give details about the main character (a 6-year-old boy who likes basketball).
ChatGPT writes the story very quickly, often in just a minute. You can use the story as it is, but it’s always a good idea to add your own thoughts to make it unique. Also, you might want to quickly check for plagiarism using free online tools, though usually, stories generated by ChatGPT are unique.
Structuring the Book with an Outline
After you have the story script, you need to plan what goes on each page. Creating a detailed outline helps you know exactly what text and images are needed for every page. This makes creating the artwork in Midjourney much simpler.
Ask ChatGPT to create a detailed outline for your story, listing what will be on each page. You can specify the maximum number of pages, for example.
Look at the outline ChatGPT gives you. You can read and change parts if needed. Sometimes, a part of the original story might be missed in the outline, especially if you limited the page count. You can either ask ChatGPT to add the missing part or handle it yourself when you create the images.
Generating Image Prompts from the Outline
The outline from ChatGPT often describes the concept for the pictures on each page. To make things easier for Midjourney, you can ask ChatGPT to turn these concepts into image prompts.
[P]Try asking: "I want to generate images for each page. Please write image prompts for these illustrations." Sometimes, you might need to explain what you want more clearly. You could even mention a tool like DALL·E (since it's also by OpenAI) to help ChatGPT understand the format needed for image generation prompts.[/H]
Review the prompts ChatGPT creates. You can use parts of these prompts for your backgrounds or as inspiration when writing your own prompts for each page in Midjourney.
Creating and Maintaining a Consistent Character in Midjourney
Now it's time to create the character images in Midjourney. First, you need to create your main character using a prompt. Include details about how you want the character to look and the style (like Pixar style). A plain background often works well for initial character images.
After generating some options, pick the character you like best and upscale it.
Midjourney has a helpful feature called Character Reference (`--cref`). This allows you to keep the same character across different images and scenes. To use it, you need the image URL of your reference character. You can get this by dragging the image into the prompt box or copying the image's link.
In your prompt for a new scene, you will include a description of what is happening on that page, followed by `--cref` and the URL of your character reference image. Like this:
`/imagine prompt [scene description] --cref [image URL] --style raw`
You can also use the `--cw` (Character Weight) parameter to control how much the new image matches the original character.
`--cw 0` focuses mostly on the character's face.
`--cw 100` (the default) includes the face, hair, and clothing.
Include all these parts in your prompt for each page’s illustration.
Fixing Small Details with Vary Region
Sometimes, details in a generated image aren't quite right, like the color of shoes or the number of fingers. After upscaling an image, you can use the Vary (Region) tool to fix these issues.
Click the Vary (Region) button. Use the selection tool (freehand or rectangular) to pick the area you want to change. Then, type what you want to see in that area and submit.
You can even use `--cref` with Vary Region. If a specific item on the character (like shoes) isn't coming out right consistently, you can cut out just that item from your original character image, upload it to Midjourney separately, and create a `/blend` or reference link for just that item. You can then use that specific item's link with `--cref` within Vary Region selections to guide Midjourney on correcting that detail.
Managing image generations and applying settings like `--cref` and `--cw` for many images can be time-consuming. A tool like the TitanXT Midjourney Automation Suite can help. It offers features like bulk prompting, preset management, and automatic variations, allowing you to generate multiple book pages faster while keeping your character consistent.
Putting it Together in Canva
Once you have the images for all your pages, you need to create the interior pages of your book. Design tools like Canva are great for this.
Open a document with the correct page size for your book (consider margins needed for printing). Upload all the images you generated in Midjourney.
Place the first image on your first page. Add a text box and paste the story text for that page (from your ChatGPT outline). You can add shapes or other elements to the page design as well.
Adjust the font style, size, and color. Combine the images and text for each page based on your outline until all the interior pages are complete.
Finishing Touches
Finally, design a cover for your book. Give it a name like "The Little Hoop Star" suggested by ChatGPT.
You can then save your pages as a file for printing or use platforms (like Simplebooklet flip app in Canva) to create a digital flipbook. You could even turn the pages into a video.
Creating a children's book involves managing many images and ensuring consistency. This is where automation can be a game-changer. The TitanXT Midjourney Automation Suite can streamline your workflow by allowing you to organize prompts, apply consistent character settings (`--cref`, `--cw`) across batches of images, and track your generated assets more effectively. Check out the suite to save time on repetitive tasks and focus more on the creative aspects of your book.
Using AI tools like ChatGPT for story creation and Midjourney's Character Reference feature for consistent illustrations makes the process of creating children's books much more accessible for authors.
Ready to make your own consistent characters for your stories? Start experimenting with Midjourney today, and consider how automation tools could help you create your next book even faster.
Explore the capabilities of the TitanXT Midjourney Automation Suite to speed up your character generation and scene creation process.






