1. Start by Creating Your Image First
This is the best practice to ensure the image is suitable and ready to be animated
Open your canvas
Right-click and select Create Image
Choose any model you prefer (Nano Banana, Imagen, Flux, etc.)
Go into Advanced Settings, set the aspect ratio (size) you need
2. Prepare the Prompt for the Image
Use a descriptive prompt, but refrain from adding any dynamics, motion, and speech elements.
Keep only:
What the scene looks like
The setting
The people
Clothing
Lighting
Style
Don't add anything about:
Movement
Speech
Camera actions
Gestures
This reduces prompt overload and gives a clean base image.
3. Generate the Image
Click Generate and wait for the image to appear
This will be the start frame of your animation.
4. Edit the Image if Needed
Double-click the generated image to open it in the editor and make adjustments before animating.
Remove unwanted objects
Replace objects
Add missing elements
Change colours
Change aesthetics
Fix defects
Apply edits until the image is exactly how you want it.
5. Avoid Letting the Video Model Create the Image
Don’t rely on the video model to generate and animate the image at the same time.
This can cause:
Extra limbs
Deformities
Inconsistent faces
Style mismatches
Using a dedicated image model first prevents these problems.
6. Prepare Your Animation Prompt
Your video prompt should now be ONLY about:
Dynamics (required motion)
Action (who moves and how)
Motion (camera behaviour)
Speech (dialogue and tone)
No visual description should be included.
The image already contains all visuals.
7. Animate the Image
Select your final image by clicking on it.
Click Animate Image
Select the model you want (e.g. Kling 2.1 Standard Video, Veo 3.1, Minimax 2.3)
Check the aspect ratio in Advanced Settings
Paste or type your short dynamic prompt
Click Generate
8. Review the Video
You should now see:
Clean characters
No extra limbs
Accurate faces
Correct movements
Clear speech timing
This works because you separated the two jobs.
9. Understand the Prompt Roles
Split your workflow into two clear parts:
A. Image Generation Prompt
Only contains:
Description (Subject, image type)
Visual details (character, background, environment)
Styles (Cyberpunk, Watercolour, Anime)
Modifiers (camera angle, lighting)
B. Video Prompt
Only contains:
Dynamics (require motion)
Action (who moves and how)
Motion (camera behaviour)
Speech (dialogue and tone)
Keeping these roles separate prevents prompt conflicts.