Skip to main content
PixVerse V6 generates video from a text prompt, animates a still image, or composes a scene from reference subjects and backgrounds. PixVerse V6 Max is the higher-fidelity tier of the same family and appears as a model option in the text to video and image to video nodes. ComfyUI runs the family through partner nodes: PixVerse V6 Text to Video, PixVerse V6 Image to Video, and PixVerse V6 Fusion (Reference to Video). They are cloud API nodes, so they need a Comfy account with credits and download no model files locally. Clips run from 1 to 15 seconds at 360p, 540p, 720p, or 1080p, with a native audio track generated by default. PixVerse V6 adds style presets and multi-shot cutting inside one generation. V6 Max renders at higher fidelity and drops the style presets. PixVerse records the seed but does not reproduce a run from it.
To use the Partner Nodes, you need to ensure that you are logged in properly and using a permitted network environment. Please refer to the Partner Nodes Overview section of the documentation to understand the specific requirements for using the Partner Nodes.
Make sure your ComfyUI is updated.Workflows in this guide can be found in the Workflow Templates. If you can’t find them in the template, your ComfyUI may be outdated.If nodes are missing when loading a workflow, possible reasons:
  1. You are not using the latest ComfyUI version (Nightly version)
  2. Some nodes failed to import at startup

​
Available workflows

​
Text to Video

Describe the shot and the model renders one clip, with native audio if you leave it enabled. This is the mode with no input files at all. PixVerse V6: Text to Video workflow preview

Run on Comfy Cloud

Open in Comfy Cloud

Download Workflow

Download JSON or search “PixVerse 6.0: Text to Video” in Template Library
Example prompts
  • Single shot, low-angle tracking across wet asphalt at night, a courier runs past neon signs, shallow depth of field, cool blue grade
  • Slow orbit around a clay teapot on a wooden table, morning window light, dust in the air, macro detail
  • Two dancers in a bright studio, handheld camera, fast footwork, natural sound of footsteps and breathing

​
Text to Video with V6 Max

The same text to video workflow on the higher-fidelity tier. V6 Max has no style presets, so the look comes from the prompt and the negative prompt. PixVerse V6 Max: Text to Video workflow preview

Run on Comfy Cloud

Open in Comfy Cloud

Download Workflow

Download JSON or search “PixVerse V6 Max: Text to Video” in Template Library

​
Image to Video

Animate a still image. The output keeps the aspect ratio of the input, so the frame shape is decided by the image rather than by a ratio setting. PixVerse V6: Image to Video workflow preview

Run on Comfy Cloud

Open in Comfy Cloud

Download Workflow

Download JSON or search “PixVerse 6.0: Image to Video” in Template Library
Input material

sunny_pink_high_tops.png

Load into the LoadImage node

​
Image to Video with V6 Max

The image to video workflow on the V6 Max tier. Multi-shot cutting is not exposed on this variant, so it produces one continuous take. PixVerse V6 Max: Image to Video workflow preview

Run on Comfy Cloud

Open in Comfy Cloud

Download Workflow

Download JSON or search “PixVerse V6 Max: Image to Video” in Template Library
Input material

pixverse_v6max_i2v_input_image.png

Load into the LoadImage node

​
Reference to Video (Fusion)

Place your own subjects into your own scene. Connect images as subjects or backgrounds and refer to each one by its slot name in the prompt, for example @Subject1 walks through @Background1. Fusion runs on PixVerse V6 only. Connecting a reference video switches the node to Omni mode: the output length then matches the longest reference video, and the video can also lend motion, framing, or style. PixVerse V6: Reference to Video workflow preview

Run on Comfy Cloud

Open in Comfy Cloud

Download Workflow

Download JSON or search “PixVerse 6.0: Reference to Video” in Template Library
Input materials

street_rider.png

Load into a LoadImage node as a subject

futuristic_urban_editorial.png

Load into a LoadImage node as a background

​
Node controls

Fusion accepts up to 7 reference images when no reference video is connected and up to 10 with one connected. Image to video keeps the aspect ratio of the input image, so aspect_ratio is not offered there.

​
Prompting tips

  • Write the shot, not the mood. One paragraph with subject, action, camera move, and light gives the model something to render: low-angle tracking shot, a courier runs past neon signs, cool blue grade. A list of adjectives alone usually drifts.
  • Describe audio when you keep it on. Footsteps, impacts, ambience, and a music style all land in the generated track. Turn generate_audio off when you plan to add your own sound.
  • Use multi_clip for a sequence. It is how a single generation returns several shots with a consistent look, and the prompt should then number the shots.
  • Label every reference in Fusion. A connected subject or background that the prompt never mentions is usually ignored, so write @Subject1 and @Background1 where they belong in the action.
  • Keep negatives short and concrete. text, watermark, logo, subtitles, oversaturated colors works better than a long list of style words.
  • Do not chase seed reproducibility. The seed is stored for reference only, so a re-run with the same seed is a new generation.

​
Get started

  1. Update ComfyUI to the latest version
  2. Go to Template Library, search for PixVerse