Prompt PixVerse for image-to-video
Last updated: October 12, 2026
PixVerse turns an approved still into a short, energetic clip, from one second up, with sound in the same pass if you want it. Pick it when the look is signed off and the job is action: a spell cast, an attack, a charged idle.
Let the agent write it
You don't have to write any of this yourself. The Layer agent knows how to prompt each model it runs, from the same guidance this page is built on. Attach the still in chat, name the model and say what should happen, for example: Animate this with PixVerse v6: the mage whips his staff forward and throws a fireball to the right, fixed camera, 1 second. The agent writes the prompt, shows you its plan, and runs it when you approve. Use the rest of this guide when you want to write or adjust the prompt yourself, and see Write a brief for the agent for how to brief it.
When to use it
Use PixVerse when you have the frame and need a burst of motion: ability casts, attacks, charged idles, short trailer beats and social clips. It packs a lot of movement into each second, so even the shortest clip holds a whole cast, and it keeps the character's design and colours from the first frame.
For slow, continuous motion and long takes, such as a breathing idle or a living menu background, see Prompt Seedance for image-to-video. To generate a clip with no image at all, see Create a video.
Versions on Layer
PixVerse v6
The version this guide's examples were made with. Clips run 1 to 15 seconds at 720p or 1080p, start from a first frame, and can generate audio in the same pass. It also takes a negative prompt.
How to write the prompt
The action and how it looks physically, then one camera instruction, then the style and what must stay the same:
The mage whips his staff forward with force, his cloak snapping back, and a fireball streaks out to the right trailing sparks and motion blur. Fixed camera. Stylised painterly game art; the character, costume and colours stay exactly as in the image.
The first words carry the most weight, so open with the action. The first frame already shows who is in the shot and where, so there is no need to describe it again: every word goes on what moves, how hard, and what the camera does. That prompt is the one in the clip above.
What works
Show speed, don't name it
Describe what speed looks like: force, cloth snapping back, a trail of sparks, motion blur. We ran "whips his staff forward with force, his cloak snapping back, and a fireball streaks out to the right trailing sparks and motion blur" against "fast swings his staff forward and a fireball fast shoots out to the right" on the same frame and seed. Both landed the cast and kept the mage on-model, but mid-swing the "fast" version smeared the arm and staff into one streak, while the physical version kept the arm, the staff tip and the fireball readable in every frame.
Motion, not description
Spend the prompt on what changes, plus one line on what must stay the same ("the character, costume and colours stay exactly as in the image"). In our test, restating the frame ("a young fire mage in a red hooded cloak with gold trim stands on a mossy stone ledge...") before the same motion caused no drift, but it added nothing either, so it is words you can drop.
One camera move
Pick one: slow push-in, pan left, tilt up, orbit, handheld tracking shot, low angle, aerial view, close-up. In our test, asking for a push-in, a pan, an orbit, a tilt and handheld shake at once got a push-in and nothing else, and a slightly less steady clip. PixVerse warns that stacked moves cause jitter and drift on longer clips.
Concrete style and light
Name the medium ("stylised painterly game art"), the light ("warm rim light from the ember") or the lens. Words like "cinematic", "epic", "8k" or "masterpiece" point at too many looks to steer anything.
Sound in the prompt
With Generate audio on, name the sounds you want ("the staff whooshes, the fireball roars, sparks crackle"). PixVerse does not reliably infer them from the action.
The same words for the same character
When you animate one character across several clips, describe them with exactly the same words every time. A synonym, such as "brunette" for "dark brown hair", reads as a different character.
What to avoid
Exclusions in the prompt
A "negative: jitter, bent limbs, flicker" list inside the prompt is read as instructions and can bring in what it names. Phrase constraints positively in the prompt, and put artefacts to keep out (flicker, distorted hands, watermark) in Negative prompt under Advanced.
Counting on a locked camera
"Fixed camera" is not a lock. On a strong action, our clips still pushed in a little. If the framing must not change, crop the result or pick a gentler action.
Complex choreography in one take
Hands and chaotic scenes with many moving parts still need retries. In our spell-cast tests the staff lost some of its length by the last frame. Plan for a few seeds on complex action, and keep each clip to one clear action.
Long prompts in other scripts
Prompts are capped at 2,048 bytes, so accented and non-Latin text fits fewer characters than English. Keep prompts to about 50 to 80 words.
Settings that matter
First frame is the still the clip starts from; the clip follows its shape. Duration runs from 1 to 15 seconds; PixVerse moves a lot per second, so find the motion at the shortest length, then render the length you need. Aspect ratio sets the resolution, 720p or 1080p. Generate audio adds sound in the same pass. Under Advanced, Negative prompt keeps artefacts out and Seed repeats a result. Above the prompt, Camera, Staging, Look and Performance take reference sets for each.
Prompts to start from
Spell cast
The mage whips his staff forward with force, his cloak snapping back, and a fireball streaks out to the right trailing sparks and motion blur. Fixed camera. Stylised painterly game art; the character, costume and colours stay exactly as in the image.
Charged idle
The mage slowly raises his staff and the ember on its tip flares brighter, shedding sparks that drift upward; his cloak stirs in a light breeze. Slow push-in, warm rim light from the ember. Stylised painterly game art; the character, costume and colours stay exactly as in the image.
Use it in Layer
Add the approved still to a session's board.
Click View on the image, then Turn into video. Create video opens in the side panel with the image as its First frame.
Click the model name at the top of the panel and choose PixVerse v6. The first frame stays in place.
In the prompt box, write the action as physical detail, one camera instruction, the style and what must stay the same.
Leave Duration at its shortest while you find the motion.
Click Create video. The cost shows on the button before you click.
When the clip lands on the board, click View to open it in the viewer and play it. Then change one thing in the prompt at a time and run it again, and render the final length once the motion is right.
Learn more
- AI video prompt guide, 7 tested fixes (PixVerse): the three-sentence prompt template, why early words carry the most weight, one camera direction per shot, and replacing "fast" and "cinematic" with concrete detail.
- PixVerse V6 review, a detailed test and guide (PixVerse): literal, physical prompting for V6, naming sounds for the native audio, and drafting cheap before rendering long.
- PixVerse V6 (PixVerse): the camera terms V6 responds to and what makes a start image animate cleanly.
- V6 model parameters (PixVerse Platform): V6's parameter ranges, durations from 1 to 15 seconds, the seed and the audio switch.
- PixVerse V6, text-to-video and image-to-video generation (Scenario): image-to-video tips with a worked prompt, and the negative prompt as the place for artefacts to keep out.