Skip to main content

How to Create AI Videos with Seedance 2.5 & 2.0

Create cinematic AI videos with director-level control — instantly and effortlessly.

In this help doc

Seedance is an AI video model that creates scenes and motion from text, with optional image, video, and audio references. This doc shows how to open Seedance, choose Seedance 2.5 or 2.0, describe your video, and set the generation options.


Open Seedance

To open Seedance, click Seedance under Generative Videos in the left menu of the dashboard. The Create Videos with Seedance 2.5 page opens.

Clicking Seedance in the left menu to open the Seedance page

Note: Seedance uses Generative Credits and is available on paid plans. Learn more about Generative Credits.

Choose a model

To choose the model, click the model name in the title (Seedance 2.5 by default) and choose Seedance 2.5 or Seedance 2.0.

  • Seedance 2.5: videos from 4 to 30 seconds. A good choice for longer scenes or several actions in a row, such as ads, UGC, and story-driven videos.

  • Seedance 2.0: videos from 4 to 15 seconds.

Opening the model list and choosing Seedance 2.5

Describe your video

To describe your video:

  1. In the box, describe the subject, action, scene, and camera work in plain language.

  2. (Optional) Click + to upload images, videos, or audio as references. Uploaded files are tagged automatically, such as @image1, @video1, and @audio1. Type @ to mention them in your prompt and tell Seedance how to use each one.

  3. (Optional) Click Multi-reference to choose how references are used: Multi-reference, Start & End, Video Editing, or Video Extension.

Typing a prompt and opening the References menu

Note: Before your first upload, you confirm the Media Upload Agreement. You must own the media or have permission to use it, and you must have consent from any person who can be identified in it.

💡 Tip: When you use several references, say clearly which element to take from each one. For more prompting tips, see the Seedance 2.5 Prompt Writing Guide.

Set the generation options

To set up the video:

  1. Choose the aspect ratio: Auto, 21:9, 16:9, 4:3, 1:1, 3:4, or 9:16.

  2. Choose the resolution: 720p or 1080p.

  3. Choose the duration, for example 10 sec. The range depends on the model.

  4. Choose With Audio to also generate dialogue, background sound, music, or sound effects, or No Audio.

  5. Check the credits next to the arrow button, then click the arrow to generate.

Setting 9:16, 10 seconds, 720p, and With Audio, then pointing to the generate button

Note: The credits change with the model, duration, and resolution. For example, at 720p Seedance 2.5 showed 10 credits for 5 seconds and 20 for 10 seconds, and Seedance 2.0 showed 8 for 5 seconds.

When the video is ready, review it. If it isn't what you expected, change your prompt or references and generate again.


FAQ

Which model should I use?
Both models support text and references. Choose Seedance 2.5 for videos up to 30 seconds or several actions in a row, and Seedance 2.0 for videos up to 15 seconds.

What happens if generation fails?
If generation fails because of a system or generation error, the credits for it are refunded automatically. Common causes are prompts that break the content policy, images smaller than 512px, and unsupported file formats.

What content isn't allowed?
Using real people without permission, violent, explicit, or hateful content, characters or brands that infringe on rights, and misleading deepfake-style content. See Why Your AI Content Request Was Blocked.

Who owns the generated video?
You do, under the AI Studios Terms of Service. You're responsible for not infringing on others' rights.

Can I edit a generated video?
Yes. Use it as a reference with Video Editing or Video Extension to change or extend the scene.

Can I run several generations at once?
You can run a limited number of tasks at the same time per account.

Is Seedance available everywhere?
It may not be available in some regions because of regional policies or infrastructure limits.

Related articles

Did this answer your question?