Reference To Video

Create videos with image, video, and audio references. Combine your materials with a prompt to guide the subject, scene, movement, and sound.

Reference materials0/50
Return Last Frame
Type @ to reference uploaded materials, e.g. @Image1 as first frame...

Generate multiple videos in one batch with the same settings, up to 10 videos.

Reference input guide
Seedance 2.5 Capability Guide

Seedance 2.5 supports text-to-video, first-frame or first-and-last-frame image-to-video, and reference-to-video with up to 50 mixed image, video, and audio materials.

Prompt Guide

Type @ to quickly insert your uploaded files into the prompt.

Select a file to insert it as a reference (e.g. @Image1, @Video1, @Audio1), so the AI knows exactly which material to use.

Material Limits
  • Up to 30 reference images
  • Up to 10 reference videos; each 2–30 seconds and 30 seconds total
  • For video editing tasks, video materials must be 4–30 seconds long, with a minimum duration of 4 seconds
  • Up to 10 reference audio clips; each 2–30 seconds and 30 seconds total
  • Maximum 50 materials in total; provide at least one image, video, or audio
  • Output: 480p or 720p, from 4 to 30 seconds
  • Text and reference modes: adaptive, 16:9, 9:16, 1:1, 4:3, 3:4, or 21:9. Image mode: adaptive only
Supported Input Combinations
  • Text + Image
  • Text + Video
  • Text + Audio
  • Text + Image + Video
  • Text + Image + Audio
  • Text + Video + Audio
  • Text + Image + Video + Audio

Audio may be used as the only reference material. Use @Image1, @Video1, or @Audio1 in the prompt to map each material clearly.

What is Reference To Video?

Reference To Video generates a new video from uploaded images, video clips, and audio files, together with your text prompt. Use images to define subjects and settings, videos to guide action and camera movement, and audio to provide sound or rhythm references. Combine these materials and explain the role of each one in your prompt. Supported formats, quantities, and combinations depend on the selected model.

Key Features of Reference To Video

Combine image, video, and audio inputs to communicate how your video should look, move, and sound.

Floating portrait, coastal landscape and car reference frames converging into one cinematic scene

Combine Image, Video, and Audio References

Upload images, video clips, and audio files in the same reference workflow. Use a portrait or product photo for the subject, a clip for movement and camera direction, and an audio file for sound or rhythm. Tell the model what to draw from each material in your prompt; available input types and limits depend on your model.

Three cinematic frames showing the same woman in a cobalt coat across different coastal locations

Carry the same character into new scenes

Reuse character images to guide facial features, hairstyle, and wardrobe while changing the location or action. Take a character from a coastal road to a city street without describing their appearance from scratch for every shot. Review each result for visual consistency.

A floating cinematic camera orbit with blue motion trails around a coastal car scene

Direct camera movement with a reference clip

Show the camera move you have in mind instead of relying on words alone. Use a supported video reference to guide an orbit, tracking shot, or slow push-in, then specify the subject and action in your prompt.

How to use Reference To Video

Upload image, video, or audio references, explain how to use them, and generate your video.

  1. Add your references

    Choose a model and upload your images, video clips, and audio files. The uploader shows the formats and limits supported by that model.

  2. Describe the scene

    Reference your uploaded materials in the prompt and explain what each should guide: the subject, setting, action, camera movement, or sound.

  3. Generate and refine

    Review the credit estimate, generate your video and adjust your direction after previewing it.

Use Cases

Create campaign videos, character scenes, and production previews using image, video, and audio references.

Product Launch Campaigns

Start with product photography, a location reference, and a clip showing the camera move you want. Build a launch-video concept that places the product in context, then compare different settings and actions with your campaign team.

Character-Based Short Films

Bring character images and scene references into your next story. Use them to guide appearance and wardrobe as you develop an introduction, a confrontation, or a closing scene, reviewing each shot for continuity before assembling the film.

Social Media Content

Combine product photos, creator images, and video or audio references to develop content for Reels, TikTok, and Shorts. Describe the opening hook, subject action, and visual style, then compare different concepts before assembling your next social post.

Dance & Action Planning

Use a rehearsal clip or movement reference alongside an image of your intended subject. Explore how the action could look from another viewpoint, helping choreographers and directors discuss staging, framing, and camera direction before production.

Architecture & Interior Presentations

Bring together room images, material references, and examples of the camera movement you want. Develop a visual presentation of a design concept, from a slow room reveal to an exterior approach, for discussion with clients.

Agency Pitches & Film Previews

Turn a client’s reference deck into a video concept before committing to a shoot. Combine casting images, location references, and sample clips to communicate the proposed composition, atmosphere, and movement more clearly than a static moodboard.

Frequently asked questions

Find answers about inputs, available controls, and the generation workflow.

Can I use image, video, and audio reference materials?

Yes. Reference To Video supports image, video, and audio reference materials with compatible models, including Seedance 2.5. You can combine supported types in one request and describe how each should guide the result. Check the selected model’s uploader for accepted formats and file limits.

Will the result exactly match my references?

References guide generation; they do not guarantee an exact match. Review identity, details and movement before using the result.

How is this different from image-to-video?

Image-to-video starts from an image. Reference-to-video can use a broader collection of materials to guide a newly generated scene.

How many credits will it use?

The generator displays the required credits before you submit. The cost depends on the selected settings and service.

Direct your next scene.

Bring your images, video clips, and audio references together, then describe the scene you want to create.

Create from references