Text to video vs image to video: which should you use?
Text to video gives freedom and image to video gives control. How to choose between them, and how to use them together for the best results.
Both modes end with a video. The difference is who decides what it looks like: with text, the model does, and with an image, you do.
Side by side
| Text to video | Image to video | |
|---|---|---|
| You provide | A description | A picture and a description of the motion |
| Control over the look | Lower | High |
| Best for | Exploring ideas | Products, characters, exact looks |
| Your prompt focuses on | Everything | Only what changes |
| Video shape | You choose | Follows the image |
When to use text
Use it when you do not yet know what you want. It is the cheapest way to try out ideas, and each run may surprise you.
When to use an image
Use it when appearance matters. Your real product, a specific character or a composition you have already approved should all start from a picture.
Use them together
- Explore with text at low resolution.
- When you like a result, save a good frame.
- Use that frame as your starting image, and refine the action.
- For several scenes with the same subject, use reference images.
Read more on text to video and image to video.
Try it while it is fresh
Try a free preview, no card needed. Videos up to 20 minutes long, with sound.
Related guides
How to Make an AI Video: A Beginner's Guide for 2026
A plain guide to making your first AI video: choosing an idea, writing the prompt, picking a length and format, fixing problems and publishing.
How to Turn a Photo Into a Video With AI
Turn any photo into a video in a few steps. Choosing a picture, writing the motion prompt, adding sound, and avoiding the problems that make it look fake.
AI Video Prompts: 30 Examples You Can Copy
30 AI video prompts you can copy for ads, YouTube, social and stories, plus a simple formula for writing your own and the mistakes to avoid.