PixVerse V4 Text to Video
PixVerse's February 2025 video model, used here to turn a text prompt into a 5- or 8-second clip at up to 1080p. Its launch post promised better realism and more natural movement.
Overview
PixVerse announced V4 on 23 February 2025 and summed it up as better realism and more natural movement. The same post introduced sound generation and Restyle, which are separate PixVerse features and not part of this page.
Choosing a variant: V4 is the middle release of the three. PixVerse V3.5 Text to Video is the December 2024 release, which PixVerse launched with a focus on anime. PixVerse V4.5 Text to Video followed in May 2025, and PixVerse said it handles complex movement more smoothly. Stay on V4 when new clips have to match a series you already started with it.
What you can build
Three of the six samples here are photorealistic: a man in a long coat walking down a dark alley, a girl laughing in a sunny park, and a firefighter walking through smoking ruins under an orange sky. Two were made with style presets: two figures in red robes beside a pair of waterfalls, drawn with the anime preset, and two motorcyclists riding toward the camera on a desert road, made with 3d_animation. Typical uses follow from that:
- B-roll and mood shots to cut into ads or explainer videos.
- Short story beats for pitch reels and storyboards.
Inputs
Only the prompt is required. Everything else has a default, so a bare prompt returns a 5-second, 16:9 clip at 540p in normal motion mode. The optional settings are:
- a negative prompt listing what should not appear;
- frame shape: 16:9, 4:3, 1:1, 3:4 or 9:16;
- length of 5 or 8 seconds;
- resolution of 360p (marked Turbo), 540p, 720p or 1080p;
- motion mode, normal by default or fast, which PixVerse permits only for 5-second clips and not at 1080p;
- a style menu (anime, 3d_animation, clay, comic, cyberpunk) that restyles the clip, as the waterfall and motorcycle samples show;
- a seed between 0 and 2147483647, or empty for a random one.
Outputs
You get a single MP4 file with no sound, sized by the resolution and aspect ratio you pick. Five of the samples here are 1280x720 and one is 1024x576, and each runs about 5.4 seconds.
Recommended settings
- Write 25 to 200 words, as PixVerse's guide advises, naming the subject, what moves, the light and the mood.
- Put camera direction in the prompt ("the camera tracks alongside", "slow push-in"), since this page has no camera setting.
- Leave style empty when the prompt describes the look, and pick a preset when you want that style applied to the whole clip.
Limitations
- The longest clip is 8 seconds, and 5 seconds at 1080p.
- A text prompt is all it accepts, and the clip comes without audio.
Safety & compliance
PixVerse runs every V4 request through content moderation. Prompts with prohibited content are rejected and return no clip. Usage is subject to the PixVerse Platform Terms of Service. According to PixVerse, its plans give no rights over other people's intellectual property, likeness, music or brands, so get clearance before generating scenes built on them.
Example prompts
Great starting points for text-to-video-v4.
API quick start
Run text-to-video-v4 with a single API call.
{
"prompt": "A dynamic, storybook-style illustration o…",
"negativePrompt": "...",
"aspectRatio": "16:9",
"duration": 5
}curl -X POST "https://api.wiro.ai/v1/Run/pixverse/text-to-video-v4" \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_WIRO_API_KEY" \
--data-binary @- <<'JSON'
{
"prompt": "A dynamic, storybook-style illustration o…",
"negativePrompt": "...",
"aspectRatio": "16:9",
"duration": 5
}
JSONOther models you might like
Explore similar video generation models.