How to Make Two or More Consistent Characters in Google Veo 3 – Complete Step-by-Step Guide

Have you ever tried making characters in AI video tools, but the results kept changing from scene to scene? One moment your character has brown hair, the next they’re blonde, and by the third scene, they look like a completely different person. I’ve been there—and it’s super frustrating, especially if you’re trying to tell a story or create an animated video where consistency really matters.

That’s why in this Mangidik guide, we’re going to learn how to make two or more characters look exactly the same across multiple scenes using Google Veo 3. This method is perfect for storytellers, YouTube content creators, and anyone making short animations. And yes, you can also use it for more than two characters if your project needs it.

The good news? We’re not going to do this the hard way. I’ve prepared an all-in-one special prompt that will help you generate consistent characters, create detailed scene visuals, and even produce short video clips—all while keeping your characters’ appearance locked in.

We’ll go step-by-step, from setting up your prompt to generating final videos. And by the end, you’ll have a smooth, professional-looking result without needing to spend hours tweaking your designs over and over again.

Step 1: Preparing the Prompt

Before you start generating images or videos, you need the perfect prompt. This is the secret sauce to keeping your characters consistent across every scene.

Here’s the all-in-one prompt template I use. You just need to fill in your details, like your characters’ descriptions, the main scene description, and how many scenes you want. Once you’ve done that, you’ll have everything you need to make consistent visuals from start to finish.

📌 INPUT: 1. Character 1 Description: [fill here] 2. Character 2 Description: [fill here] 3. Main Scene Description: [fill here] 4. Number of Scenes: [fill here] ---------------------------- 1️⃣ TEXT-TO-IMAGE PROMPT: CHARACTER 1 ---------------------------- Ultra realistic, full-body character design for animation/video. [Character 1 Description], with detailed facial features, expressive eyes, natural skin texture, realistic hair strands, wearing [describe clothing], standing in a neutral pose, soft studio lighting, plain neutral background, highly detailed textures, consistent with animation-ready style, cinematic color grading. ---------------------------- 2️⃣ TEXT-TO-IMAGE PROMPT: CHARACTER 2 ---------------------------- Ultra realistic, full-body character design for animation/video. [Character 2 Description], with detailed facial features, expressive eyes, natural skin texture, realistic hair strands, wearing [describe clothing], standing in a neutral pose, soft studio lighting, plain neutral background, highly detailed textures, consistent with animation-ready style, cinematic color grading. ---------------------------- 3️⃣ SCENE BREAKDOWN ---------------------------- Scene 1: [Describe moment 1 in detail, including setting, actions, and expressions] Scene 2: [Describe moment 2 in detail] Scene 3: [Continue until reaching Number of Scenes] ---------------------------- 4️⃣ SCENE PROMPTS ---------------------------- --- SCENE 1 --- a. TEXT-TO-IMAGE SCENE PROMPT Ultra realistic, [Detailed scene description based on Scene 1], both characters in consistent visual style with earlier designs, realistic lighting, natural shadows, cinematic framing, detailed background matching the setting, animation-ready style. b. IMAGE-TO-VIDEO SCENE PROMPT (Veo 3) Ultra realistic, animated sequence based on the above scene description, smooth camera movement (slow dolly or pan), 5–7 seconds duration, soft cinematic transitions, realistic lighting, characters moving lips naturally as if speaking (lip-sync only, no sound). Dialogue (for lip-sync reference only, no sound in output): Character 1: "[Line from Character 1 matching the scene]" Character 2: "[Line from Character 2 matching the scene]" --- SCENE 2 --- a. TEXT-TO-IMAGE SCENE PROMPT Ultra realistic, [Detailed scene description based on Scene 2], both characters in consistent visual style with earlier designs, realistic lighting, natural shadows, cinematic framing, detailed background matching the setting, animation-ready style. b. IMAGE-TO-VIDEO SCENE PROMPT (Veo 3) Ultra realistic, animated sequence based on the above scene description, smooth camera movement (slow dolly or pan), 5–7 seconds duration, soft cinematic transitions, realistic lighting, characters moving lips naturally as if speaking (lip-sync only, no sound). Dialogue (for lip-sync reference only, no sound in output): Character 1: "[Line from Character 1 matching the scene]" Character 2: "[Line from Character 2 matching the scene]" [Repeat the same structure for all scenes] 

Once you’ve filled in your characters and scene details:

1. Copy the whole prompt.
2. Open ChatGPT.
3. Paste the prompt.
4. ChatGPT will return fully detailed prompts for both characters, all scenes, and ready-to-use image-to-video instructions.

Step 2: Generate Character Images

Now that you have your prompts, it’s time to create your characters.

1. Copy the text-to-image prompt for Character 1 from ChatGPT.
2. Open your favorite text-to-image generator. Personally, I use Flux because it’s stable and gives great details.
3. Paste the prompt into Flux.
4. Set the aspect ratio to 16:9 and keep the batch size to 1 image (to save tokens or credits).
5. Hit Send and wait a few seconds.
6. Your first character is ready!
7. Repeat the same process for Character 2 using their prompt.

The key here is to save these images—you’ll be uploading them later when generating your scenes.

Step 3: Generate Scene Images

This step is where your characters come to life in specific settings.

1. Copy the text-to-image prompt for Scene 1 from ChatGPT.
2. Go back to Flux.
3. Click the image icon inside the chat box and upload your Character 1 image.
4. Click the image icon again to upload your Character 2 image.
5. Paste the scene prompt you copied earlier.
6. Again, set batch size to 1 and aspect ratio to 16:9.
7. Tap Send and wait for your scene image to generate.
8. Download it and repeat for all scenes.

By uploading both character images every time, you’re telling the AI exactly how they should look—keeping them consistent across all scenes.

Step 4: Generate Videos

Now comes the fun part—turning those scene images into short animated clips.

1. Copy the image-to-video prompt for Scene 1 from ChatGPT.
2. Go to the Google Gemini website.
3. Select the Video option in the chat box.
4. Upload your Scene 1 image.
5. Paste the prompt.
6. Hit Send and wait a few minutes for it to process.
7. Download your video clip.
8. Repeat for every scene.

This method gives you 5–7 second video clips with smooth camera movements, cinematic lighting, and natural lip-sync animations (even though there’s no audio).

Step 5: Preview Your Work

Once you have all your scene videos, put them together in your video editor of choice. Check if the lighting, colors, and character appearances match across the entire project. If you see small differences, you can correct them with light editing—but in most cases, this prompt method will keep everything consistent.

Video Tutorial Reference

If you prefer to follow along with a visual guide, I’ve made a video tutorial here:

Conclusion

Creating consistent characters in AI-generated videos can feel like an impossible task, but with the right prompt structure and a clear workflow, it’s actually pretty simple. By breaking the process into steps—prompt preparation, character generation, scene creation, and video animation—you can make professional, story-driven clips without the usual frustration of “character drift.”

Whether you’re making short films, educational videos, or animated stories for social media, this approach will save you time and give your project a polished, cohesive look. And the best part? Once you’ve built your prompt once, you can reuse it for future projects—just swap in new character descriptions and scenes.

So, grab the prompt, fire up Flux and Google Veo 3, and start creating your own cinematic story today.