How to Access Seedance 2.0: AI Video Generation

You can access Seedance 2.0 AI video generation in four places as of October 2026: Dreamina, ByteDance's official web app; CapCut, for paid users; developer APIs such as fal; and VideoGen, where Seedance 2.0 Mini and Seedance 2.5 clips sit inside a full editing project. The right route depends on whether you want a single clip, an API, or a finished video.

This guide covers each route. It then walks through using Seedance AI inside VideoGen without a ByteDance, Dreamina, or fal account.

What Seedance 2.0 is

Seedance 2.0 is ByteDance's multimodal video model, released in February 2026. It takes text, images, video, and audio as inputs and generates clips up to 15 seconds long. Stereo audio and lip-sync come out in the same pass as the picture, so a character can speak a line without a separate dubbing step.

The model ranks in the top three of the Artificial Analysis Text-to-Video Leaderboard, as listed on the VideoGen Seedance 2.0 page. The Seedance overview compares it with Seedance 2.5, which handles longer clips and more reference material in one generation.

Four ways to access Seedance 2.0

Route	Who it fits	Account needed	What you get
Dreamina	Individuals testing single clips	Dreamina account	Free daily credits and up to 12 references per generation (9 images, 3 videos, 3 audio clips)
CapCut	CapCut subscribers who already edit there	Paid CapCut plan	"Dreamina Seedance 2.0" inside AI Video and Video Studio, by region
Developer APIs (fal and others)	Developers building their own product	API account with billing	Programmatic access, including Seedance 2.0 Mini endpoints
VideoGen	Teams making a complete video	VideoGen account	Seedance clips in a timeline with voiceover, captions, music, and other models

Some detail on each route:

• Dreamina is ByteDance's international consumer app and the official web home of Seedance 2.0. Jimeng is the Chinese app.
• CapCut added Dreamina Seedance 2.0 to AI Video and Video Studio for paid users in phased regional rollouts starting March 2026. The US was added on April 7, 2026.
• Developers can call Seedance 2.0 through fal and other API hosts. Fal's Seedance 2.0 prompting guide covers prompt structure for its endpoints. BytePlus ModelArk, ByteDance's first-party international API, states that its services are not available in the United States.
• VideoGen includes Seedance 2.0 Mini and Seedance 2.5 in its video clip quality tiers. AI video clip generation needs a paid VideoGen plan; the free plan includes limited AI credits for other AI features and does not include exports. See pricing.

How to use Seedance 2.0 in VideoGen

VideoGen runs in a desktop browser. A Seedance 2.0 Mini clip in VideoGen can run from 4 to 15 seconds and a Seedance 2.5 clip from 4 to 30 seconds; VideoGen never shortens a request to fit a model.

1. Start a new project, or use the prompt box on the Seedance 2.0 model page. That prompt box carries your idea into VideoGen. It does not lock the request to one model.
2. Choose a workflow. Storyboard-to-video fits a planned sequence of shots, and in a storyboard scene the spoken script may fill the whole clip. Script-to-video fits a narrated video. You can also generate a single clip from the editor.
3. Pick a quality tier where Seedance is listed. The model chooser shows a "Powered by" line for each tier. As of October 2026, every tier lists Seedance 2.0 Mini and Seedance 2.5 among its models. VideoGen picks a model per request from that tier. Seedance 2.5 is the first model tried on the Max tier; on High, Seedance 2.5 is used when Seedance 2.0 Mini cannot keep the requested length.
4. Write the prompt in five layers: subject and action, environment and lighting, camera and framing, style and mood, audio and dialogue. Put spoken lines in quotation marks and describe how they are delivered.
5. Add references if the prompt is not enough. Give each one a role, such as "use image 1 for the product shape."
6. Review the clip. Check the motion, subject consistency, audio sync, and whether the last frame gives you a clean edit point. You can regenerate one scene without touching the rest.
7. Finish the video. Add an AI voiceover, enable captions across the project in one step, pick or generate background music, and export.

For prompt-writing detail, read how to use Seedance 2.0.

A first prompt to try

Copy this prompt into a storyboard scene or a single clip request:

A barista pours latte art in a small, sunlit cafe. Steam rises from the cup as the camera pushes in from a medium shot to a close-up of the pour. Warm natural light, soft film grain. The espresso machine hisses in the background. The barista looks up and says, "Your coffee is ready," with a quick smile.

The prompt gives the model one physical action, a setting, a camera move, a style, and a short spoken line. Keep dialogue short so the mouth movement holds inside a 15-second clip. If the subject drifts, add a reference image or state what must stay unchanged.

The Seedance prompts page has more copy-paste prompts for action, dialogue, product ads, and multi-shot stories.

Which route to pick

• Pick Dreamina when you want to test the model on single clips and use many references per generation.
• Pick CapCut when you already pay for CapCut and want Seedance inside that editor.
• Pick an API host such as fal when you are a developer building Seedance into your own product.
• Pick VideoGen when the clip is one piece of a longer video. You get the same model family beside voiceover, captions, music, stock footage, and other video models, with one timeline to edit all of it.

For the kinds of videos where Seedance 2.0 does its best work, see where to use Seedance 2.0.

Frequently asked questions

Do I need a ByteDance or Dreamina account to use Seedance 2.0 in VideoGen?

No. VideoGen includes Seedance 2.0 Mini and Seedance 2.5 in its video clip quality tiers, so a VideoGen account is enough. AI video clip generation needs a paid VideoGen plan.

Can I use Seedance 2.0 in CapCut on a free plan, or only on a paid one?

As of October 2026, CapCut lists Dreamina Seedance 2.0 inside AI Video and Video Studio for paid users. The rollout is by region, with the US added on April 7, 2026.

What happens if I ask VideoGen for a 20-second Seedance 2.0 clip?

Seedance 2.0 Mini accepts 4 to 15 seconds in VideoGen. VideoGen does not shorten the request; it moves it to a model that supports 20 seconds, such as Seedance 2.5, which accepts 4 to 30 seconds.

Can I pick Seedance 2.0 by name in VideoGen?

No. VideoGen routes each request to a model from the quality tier you choose, and the "Powered by" line shows which models that tier can use. A different provider can be used as a fallback for the same inputs.

Start a project and generate your first Seedance clip in a timeline you can finish.

Sources

• Seedance 2.0 on VideoGen
• Seedance model family on VideoGen
• Dreamina: Seedance 2.0
• CapCut newsroom: Dreamina Seedance 2.0 rollout
• Fal Seedance 2.0 Prompting Guide
• VideoGen pricing
• VideoGen blog: How to Use Seedance 2.0: Prompting and Multi-Shot Video
• Seedance Prompts on VideoGen
• VideoGen blog: Where to Use Seedance 2.0: Best Use Cases