Nano Banana 2 & Pro Sale — 15% OFF | Apr 1–15 Only
Inicio/Explorar/Vidu Models/vidu/q3/reference-to-video
image-to-video

image-to-video

Vidu Q3 Reference-to-Video Mix

vidu/q3/reference-to-video

Vidu Q3 Reference-to-Video Mix generates multi-entity consistent videos from 1-4 reference images with text prompt guidance. Supports 360p to 1080p resolutions, up to 16 seconds duration, multiple aspect ratios, and optional audio generation. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Input

Drag & drop or click to upload

preview

Idle

Tu solicitud costará $0.35 por ejecución.

Con $10 puedes ejecutar este modelo aproximadamente 28 veces.

Una cosa más:

EjemplosVer todo

README

Vidu Q3 Reference-to-Video

Vidu Q3 Reference-to-Video generates cinematic video guided by multiple reference images and a text prompt. Upload one or more reference images to establish character identity, style, or scene elements — the model produces a cohesive, motion-consistent video that blends all provided references into a new scene.

Why Choose This?

  • Multi-image reference generation Upload multiple reference images to combine characters, styles, or visual elements from different sources into a single coherent video.

  • Character-consistent output The model preserves the identity, appearance, and style of subjects from your reference images throughout the generated clip.

  • Optional audio generation Enable generate_audio to produce synchronized ambient sound and atmosphere alongside the video.

  • Flexible resolution options Generate at 480p, 720p, or 1080p to match your quality and budget requirements.

  • Reproducible results Use the seed parameter to lock in a specific output for exact reproduction.

  • Prompt Enhancer Built-in tool to automatically improve your scene descriptions for richer output.

Parameters

ParameterRequiredDescription
promptYesText description of the scene, action, and atmosphere.
imagesYesOne or more reference images. Click Add Item to include additional images.
aspect_ratioNoOutput aspect ratio. Default: 16:9.
resolutionNoOutput resolution: 480p, 720p (default), or 1080p.
durationNoClip length in seconds. Default: 5.
generate_audioNoWhether to generate synchronized audio for the video. Default: enabled.
seedNoRandom seed for reproducible results.

How to Use

  1. Write your prompt — describe the scene, characters, camera movement, and atmosphere. Use the Prompt Enhancer for better results.
  2. Upload your reference images — provide one or more images via URL or drag-and-drop. Click Add Item to add more.
  3. Select aspect ratio — choose the format that fits your target platform.
  4. Select resolution — 480p for lower cost, 720p or 1080p for higher quality.
  5. Set duration — choose your desired clip length in seconds.
  6. Enable audio (optional) — generate_audio is enabled by default; disable it if you only need video.
  7. Set seed (optional) — fix the seed to reproduce a specific result in future runs.
  8. Submit — generate, preview, and download your video.

Pricing

Duration480p720p / 1080p
5s$0.35$0.77
10s$0.70$1.54
15s$1.05$2.31

Billing Rules

  • 480p: $0.07 per second
  • 720p / 1080p: $0.154 per second (2.2× base rate)

Best Use Cases

  • Character-Driven Storytelling — Place consistent characters from reference images into entirely new scenes.
  • Social Media Content — Produce short-form video clips with consistent visual identity from reference photos.
  • Marketing & Brand Video — Generate new scenes featuring consistent brand characters or products from reference imagery.
  • Creative Concepting — Rapidly prototype multi-character or multi-element scenes for pitching and storyboarding.
  • Style-Consistent Series — Maintain a unified visual style across multiple video generations using the same reference set.

Pro Tips

  • Use clear, well-lit reference images with distinct subjects for the most accurate identity preservation.
  • The more reference images you provide, the richer the model's understanding of your intended style and characters.
  • Disable generate_audio if you plan to add your own soundtrack in post-production.
  • Fix the seed once you find a result you like to iterate consistently across resolution and duration changes.
  • Use 480p to rapidly test your prompt and reference combination before committing to a higher-resolution render.

Notes

  • Both prompt and images are required fields; all other parameters are optional.
  • generate_audio is enabled by default — disable it explicitly if you do not need audio.
  • Ensure image URLs are publicly accessible if using links rather than direct uploads.
  • Please ensure your content complies with Vidu's usage policies.

Related Models