One of the challenges in AI image generation is keeping characters consistent across multiple images and poses. ConsistentCharacter.ai is a new tool that aims to address that challenge by focusing specifically on generating character images. While their website shows mostly Pixar-style animated characters, my initial experiments with photorealistic characters were reasonably effective. The tool had some glitches, but this is something I’d consider using for generating characters for scenarios.
This is not a paid review. I requested and received a limited number of credits to test out the tool, but I didn’t mention my blog in that request. These are my honest opinions and reactions from testing the tool.
Update 4/23/25: Consistent Character.ai has announced that they will stop offering photographic style characters in May 2025 to focus on cartoon and illustrated styles. My examples below focus on photographic characters for training, so nothing like these examples will be possible anymore. However, if you’re looking for illustrated and cartoon style characters, this is still a good option. If you want photorealistic characters, Midjourney creates the most realistic characters.
ConsistentCharacter.ai options
The best place to start is with their Step-by-Step guide, which includes video tutorials and prompting tips.
- Prompt Easy: Give it a basic prompt and it will enhance it for you. You can also provide an image to generate a prompt.
- Character Turbo: Provide a prompt and pick a style. You can also note actions, expressions, and background.
- Create Avatar: Create a character based on a reference photo
- Expression Editor: Change the facial expression of an existing image with detailed controls
Creating a character
First, I tried creating a photographic character and generated four poses. I was fairly happy with the results. Her clothes and hair are consistent. Her facial structure looks pretty close in all of the images. She looks a little too much like a model for my preferences, especially in the first pose, but I could fix that with better prompting.
Turbo Character
The character generated in Turbo Character was more realistic. Once you generate this, you can remove the background or edit the expression. However, I didn’t see an easy way to put this character in different poses after the Turbo Character generation. I think that feature may exist, but I couldn’t find it.
Expression Editor
In the Expression Editor, you can fine tune the facial expressions in a reference image. You can adjust the head position and head tilt, eye position, eyebrows, blink, wink, and mouth shape. The image below shows all 3 tabs for editing expressions side by side so you can see it all at once.
For this test, I tried starting with an image I had previously generated in Midjourney. This was my reference image.
The controls for editing the expression work; I love the head tilt. The controls would definitely require some practice to get familiar with them and get better results though. As you can see in the images below, my first attempt at a frowning face created a really exaggerated, unrealistic expression. My second adjustment was better, and I think with additional time and practice I could have done more. The expression editor is the feature that I feel most strongly differentiates this tool from other available options.
A few issues
One glitch I ran into is that when you use the Expression Editor, it tends to remove your reference image after you make a change. I guess that’s so you can upload a new image, but in practice, it feels like a glitch. I think you’d normally want to make multiple poses of the same reference image.
There also doesn’t seem to be a way to go back and see which settings I used to generate previous expressions. I wish I had that ability to see it so I could learn how to use the tool better.
This tool is pretty recently out of beta though, and they’re making improvements regularly. I expect it will improve in time.
The lowest tier paid plan is currently $29/month, which is comparable to what I have paid for character image libraries in the past.
More on consistent characters
- Diana Zdybel, the co-founder of ConsistentCharacter.ai, recently shared a comparison of AI video animation tools using an identical character image from ConsistentCharacter.ai as an animation reference.
- For a comparison of the process in another tool, check out my post Generating Consistent Characters in the Midjourney Web Interface.
- Another option for realistic characters is Ingrasp, which was specifically designed for characters for training.
Creating consistent characters is a big need for entertainment, marketing, and other fields. A lot of companies are investing in building tools that can do this for those other fields, so we in L&D will benefit from that work. In my own experiments, I haven’t found that tools based on face swapping to be particularly effective for what I need. A number of tools also feel like they might be useful in production in 6-12 months, but not today. They’re getting closer though, and I think we’ll have multiple options by this time next year.
I’ve also seen previews of another tool that is geared specifically towards elearning and workplace training scenarios that looks really promising. I’m looking forward to doing more experiments with that tool. Stay tuned for more!
Upcoming events
Instructional Design Uncovered: A Community Conversation with Instructional Designers. September 10. ATD Research Triangle Area
Creating High-Quality AI Media: Prompting, Scripts, and Consistency. October 8. Register for the ATD AI Intensive.
Creative Solutions with Twine: Design, Streamline, and Refine Branching Scenarios. October 14. ATD Dallas Learning Technology SIG
DevLearn
Register for DevLearn by October 2 for the Advanced Rate. Save an extra $100 with code CHRISTY.
Branch Smarter, Not Harder: Naming Conventions, Variables, and Bottlenecks in Storyline
- November 3, 2026 at 9:20 AM
- CE01: Storyline & Rise Pro Lab
Cohesive by Design: Generate Consistent AI Icons, Illustrations & Characters
- November 4, 2026 at 10:00 AM
- BYOD (Bring Your Own Device)
Buzz Discussion: Video & Media
- November 4, 2026 at 3:45 PM
Wow, ConsistentCharacter.ai sounds amazing! Getting multiple poses and an Expression Editor for character generation is seriously next-level. It makes me think about all the cool AI image tools out there. Like, if you ever need to quickly clean up an image or remove something, AI Remove Text can be a real lifesaver!
This sounds awesome! An Expression Editor for character images is a game-changer. It highlights how important consistency is, not just visually, but for their whole persona. That’s why a good Consistent Character AI Generator is so key for all those details. So cool!
This looks like an extremely helpful tool for e-learning, so thanks for sharing.
From what I’m seeing in the market, I think we’ll have 3-5 of these sorts of consistent character generation tools good enough to use for at least some production by the end of the year. I’m still mostly using Midjourney, but this is definitely an option. InGrasp.ai looks promising as well, especially since that’s specifically intended for training images and not game development or other purposes.