How to Make a Consistent Character Video Using Google Veo 3 (Step-by-Step)

July 21, 2025 Post a Comment

Hey there! In this Mangidik article, I’m going to show you exactly how to make a video with a consistent character using Google Veo 3—and yes, it’s easier than you think. No more worrying about AI randomly changing your character's face every time. With this guide, you’ll be able to create story-style videos with the same character across multiple scenes.

Let’s jump right in!

Step 1: Preparing the Prompt

Before we get into any AI tools or video generation, the very first thing you need is a prompt. But not just any prompt—we’re talking about three specific types:

A prompt to design your character (text-to-image).
Prompts to create each scene your character will appear in (also text-to-image).
Prompts to turn those scenes into videos (image-to-video for Google Veo 3).

Now, I’ve made your life super easy. I’ve already prepared a ready-to-use ChatGPT prompt that will automatically give you all three of these types. No complicated editing or prompt engineering required.

I want you to create a character description based on the following two things:
1. Character name: Linda
2. Character profession or role: Zoo keeper
Provide the following output:
1. Complete text-to-image prompt for the character as an initial visual reference, including:
- Description of physical characteristics
- Clothing style appropriate to the profession
- Accessories, distinctive props
- Facial expressions, poses
- Voice or speaking style if featured in a video
- Pre-made prompt format for AI Image Generator (e.g., MidJourney, Leonardo, etc.)
2. Create a text-to-image prompt per scene using the reference character. Number of scenes: 3
3. For each scene, create a Google Veo 3 image-to-video prompt that includes:
- 16:9 resolution
- Description of the surrounding atmosphere
- Camera movement (pan, tracking, zoom in, etc.)
- Spoken lines as if the character were speaking to the camera
- Visual details (lighting, depth of field, color, etc.)
Compose into code blocks for easy copying. Separate character prompts, while character reference prompts and image to video are combined but per scene.

All you need to do is fill in three things:

The name of your character (for example: Linda)
The role or profession of your character (like: zookeeper)
The number of scenes you want (tip: stick to 3, since Veo 3 currently only allows 3 videos per day)

Once you’ve filled it out, just:

Copy the prompt you’ve edited
Head over to ChatGPT.com
Paste it into the chat and hit Send
ChatGPT will respond with your character prompt + all scene prompts + Veo video prompts

Pretty cool, right?

Step 2: Designing Your Character

Okay, now it’s time to bring your character to life with an image.

Take the character design prompt ChatGPT gave you.
Open the AI image generator called Flux Kontext Pro (you get 200 free credits when you sign up).
Paste the character prompt into the prompt box.
Tap the three-dot menu and set the aspect ratio to 16:9 so it’s perfect for video.
Set the batch size to 1 (this saves your credits).
Tap Send and give it a few seconds.
Boom! You’ve got a character image. Download it and save it—we’ll need it soon.

Pro tip: Make sure the character is clearly visible and in a pose that looks natural. This will help a lot in the next steps.

Step 3: Creating Scene Images (Scene by Scene)

Now that you’ve got your character image, it’s time to build the actual scenes your video will show.

Here’s how:

Copy Scene 1's prompt from ChatGPT.
Go back to the Flux Kontext homepage.
Upload the character image you created earlier (this helps keep the character consistent).
Again, make sure batch size is set to 1.
Paste the scene prompt into the prompt box.
Tap Send, and wait a few seconds.
Once the image is ready, download it.

Repeat this process for Scene 2 and Scene 3 using their respective prompts from ChatGPT.

And just like that—you’ve now got all your scenes with your consistent character!

Step 4: Generating Videos with Google Veo 3

Alright, here comes the exciting part—turning your images into moving video using Google Veo 3 (via Gemini).

Here’s what to do:

Open Google Gemini.
Tap the Video option inside the chat box.
Upload the Scene 1 image you just created.
Copy the image-to-video prompt for Scene 1 from ChatGPT.
Paste that prompt into Gemini’s chat.
Hit Send and wait about 2–3 minutes.

Voilà! You’ve got your first video.

Repeat the same for Scene 2 and Scene 3. Just change the uploaded image and prompt each time.

When you’re done, download all the video clips. You now have a three-part video featuring the same consistent character across all scenes. It’s storytelling magic—AI style.

Editing and Final Touches

Once you’ve downloaded your videos, you might want to add music, subtitles, transitions, or other edits. You can easily do this in apps like:

CapCut
KineMaster
Spring
VN Editor
Or any other video editor you like

Feel free to trim the clips, adjust the speed, or add effects. You’re the director now!

And if you want to see what the final result looks like, check out my YouTube tutorial right here:

Final Thoughts

Making a video with a consistent character in Google Veo 3 is totally possible—and honestly, kind of fun. Here’s a quick recap of the process:

Start by generating smart prompts with ChatGPT (I’ve already prepped it for you)
Design your character using Flux Kontext Pro
Create scene images using the character as a visual reference
Turn those images into video clips using Google Veo 3
Edit and polish your final videos however you like

With just a few simple tools and a bit of creativity, you can build short films, animations, or storytelling videos that look incredibly professional—even if you’re doing it all from your phone.

Got questions? Feel free to leave a comment on the YouTube video or check out the full list of links and resources on my website. Now go ahead—bring your characters to life and make something amazing.