A still image is where most people stop. Here's how to make her actually move — the image-to-video workflow, which pictures animate best, and what to realistically expect.
You've made a picture you love. She's just… standing there. The obvious next thought: can I make her move?
You can. It's called image-to-video, and it's the single biggest upgrade from "nice picture" to "wait, that's kind of real." Here's exactly how it works and how to not waste credits on it.
Text-to-video (typing a prompt and getting a clip from nothing) is still messy and expensive. Image-to-video is different and much better for this: you give it a still image you already like, and the model animates that exact frame — a few seconds of subtle motion, breathing, hair, a shift in pose. Because it starts from your image, she keeps her face and her look. You're not rolling the dice on a new person.
Short version: generate the image first, then bring it to life.
slow turn toward camera, soft smile). Keep it simple — you're describing movement, not a whole new scene.That's the whole loop: still → motion, no video-editing skills required.
Not every picture makes a good clip. From a lot of trial and error:
| Works well | Struggles | |---|---| | Clear single subject, facing camera | Busy scenes with lots of small detail | | A bit of empty space around her | Extreme close-ups (nowhere for motion to go) | | Simple pose that could move naturally | Complex hand positions (hands are still the #1 tell) | | Good lighting, sharp face | Already-blurry or low-quality stills |
Pick a clean, well-lit image and the motion looks natural. Feed it a chaotic one and you get warping.
A saved image is a moment. A short clip is presence — it's the difference between a photo and her looking back at you. Pair it with a consistent character and you've got the same girl, in motion, whenever you want.