Every app turns a photo into video now, but most make you pay, and the results melt because the prompt is wrong. This is free, and it does the hard part: the prompt. Free, no signup, about 30 seconds. Upload any photo, say what should move, and it writes a clean image-to-video prompt using the one rule that beats everything: describe only the NEW motion, not the scene that's already in the shot. Paste the prompt into Seedance or any AI model and render.

Hey Product Hunt 👋 I'm Guillermo. "Image to video" is everywhere now, but something kept bugging me: most tools make you write the prompt, and when you do, the clip comes out warped and melting. People blame the model. It's almost always the prompt. There's one rule that fixes most of it: describe ONLY the new motion you want, not the scene that's already in the photo. The model already has the picture. If you re-describe the whole scene, it fights your image and everything drifts. Tell it just the motion (a slow push-in, hair moving in the wind, steam rising) and it stays locked to your shot. So this tool does that for you, for free. Upload any photo, say what should move, and in about 30 seconds it writes a clean image-to-video prompt from your reference keyframe. You paste it into Seedance or any AI video model and render the clip yourself. No signup, no app. Just the prompt that actually works. What photo are you going to bring to life first? 📸

does it work with photos that have a lot going on in the background, or is it better when the subject is clearly separated from everything else?

Does the prompt stay grounded in the uploaded image or does it just describe the motion without any reference to the actual content of the photo?

The “only describe the new motion” tip actually made my Seedance renders look way less melted. Took my dog photo and got a clean tail-wag prompt in under a minute, no signup.

Does the prompt stay consistent if I upload the same image again, or does it generate a different one each time?

does it work with photos that have a lot going on in the background, or is it better when the subject is clearly separated from everything else?

Does the prompt stay grounded in the uploaded image or does it just describe the motion without any reference to the actual content of the photo?

The “only describe the new motion” tip actually made my Seedance renders look way less melted. Took my dog photo and got a clean tail-wag prompt in under a minute, no signup.

Does the prompt stay consistent if I upload the same image again, or does it generate a different one each time?

Image To AI Video Prompt Generator

Turn any photo into an AI video prompt in 30 seconds

Turn any photo into an AI video prompt in 30 seconds