Skip to main content

What Is Seedance? How to Use ByteDance's AI Video Model

Updated 2026-10-13

On this page

Seedance is a family of AI video models made by ByteDance Seed. It makes short videos with sound from text, images, audio or video. To use Seedance, pick a ByteDance app, an API platform or a ready-made template. The newest version in this guide is Seedance 2.5, which TechNode reported on July 31, 2026.

If you want the fastest way to try it, use a ready-made template. At Effect Farm you upload one photo, pick a template and get a video in 3–6 minutes. The templates on effectfarm.com run on Seedance.

Key facts

  • Maker: ByteDance Seed, a team inside ByteDance.
  • Type: AI video model. It makes video and audio together.
  • Seedance 2.0: announced February 12, 2026. Up to 15 seconds per clip. Up to 9 images, 3 video clips and 3 audio clips as references.
  • Seedance 2.5: reported on July 31, 2026. Up to 30 seconds per clip. Up to 30 images, 10 videos and 10 audio clips as references.
  • Where to use it: ByteDance apps, API platforms and tools built on the model. Prices and limits differ by platform.
  • Fastest start with one photo: a template page such as Make Picture Dance. One video costs 30 credits ($4.99 in the smallest pack).
  • Last checked: October 13, 2026. Sources are at the end of this page.

What is Seedance?

Seedance is a video generation model. A video generation model takes a request and returns a clip. The request can be a text prompt, a photo, a sound file or a short video.

The model comes from ByteDance Seed, a team inside ByteDance.

The official Seedance 2.0 page says the model uses a "unified multimodal audio-video joint generation architecture" [6]. In plain words, one model handles the picture and the sound in the same pass. It supports text, image, audio and video as inputs.

People search for "Seedance AI", "Seedance 2.0" and "Seedance 2.5". All three names point to the same family of models. The number tells you the version.

Seedance versions

Seedance has had several releases. The table lists the ones with an official or well-sourced description.

VersionWhat we can confirmSource
Seedance 1.0Official page says it can create 1080p videos and supports multi-shot video from text and from an image. Release date not confirmed here.[1]
Seedance 1.5 ProOfficial page says it generates voices and spatial sound effects that match the picture, with lip-sync in several languages.[2]
Seedance 2.0Announced February 12, 2026. Up to 15 seconds, multi-shot, two-channel stereo audio.[3]
Seedance 2.5Reported July 31, 2026. Up to 30 seconds in one generation, with the option to extend twice.[4] [5]

A full comparison of the two newest versions is in Seedance 2.5 vs Seedance 2.0.

What Seedance is good at

ByteDance describes the strengths of Seedance 2.0 in its launch post [3]. We list them here in plain words.

  • Motion that holds together. ByteDance says it works on complex movement and on scenes with more than one subject.
  • Reference input. You can give the model images, video clips and audio clips. It can follow the camera movement, the framing and the sound of a reference.
  • Sound with the picture. The model makes background music, ambient sound and voices as separate tracks in a two-channel mix.
  • Multi-shot clips. One clip can hold more than one shot.
  • Editing and extension. The launch post lists video editing and video extension as features.

The same post names weak spots [3]. ByteDance lists detail stability, text on screen, consistency across several subjects, complex edits and occasional audio distortion. Treat these as limits when you plan a video. Keep one clear subject and avoid text in the picture.

Seedance 2.5 adds a longer clip (30 seconds in one generation) and, according to ByteDance, smoother motion and more precise use of reference videos [4]. The details are in the version comparison.

How to use Seedance

There are three ways to use Seedance. Pick the one that matches how much control you want.

  1. A ByteDance app. TechNode reports that ByteDance is rolling Seedance 2.5 out to the Jimeng AI app and to the Pro version of Doubao [5]. Features and regions can differ.
  2. An API or a creative platform. Developers and teams call the model through an API. Third-party platforms publish Seedance 2.0 and 2.5 with their own prices, limits and sign-up rules [7] [8].
  3. A ready-made template. You do not write a prompt or choose settings. You upload a photo and pick a result such as a dance. This is how Effect Farm works.

Is Seedance 2.0 free?

It depends on the platform. The ByteDance pages that we read give no price for Seedance 2.0. Each app or API platform sets its own price, trial rules and limits, and they change. Check the platform before you start. On Effect Farm, one video costs 30 credits ($4.99 in the smallest pack).

How to use Seedance for a dance video from one photo

This is the shortest path. It needs one photo and no prompt.

  1. Pick a template. Make Picture Dance and Genjutsu AI turn one photo into a dance video.
  2. Upload one clear photo of yourself. The photo must show you, and you must be 18 or older.
  3. Pay for the video. One video costs 30 credits. See pricing.
  4. Wait 3–6 minutes. You get a link to your video by email.

If you want to write your own prompts, read the Seedance prompt guide. If you want the full photo-to-dance tutorial, read how to make an AI dance video from one photo.

Inputs, outputs and limits

Limits change with the version and with the platform. The table shows what the official pages and one API platform list. API platforms can set tighter limits than the model itself.

ItemSeedance 2.0Seedance 2.5
Clip lengthUp to 15 seconds (ByteDance)Up to 30 seconds in one generation (ByteDance)
Reference imagesUp to 9 (ByteDance)Up to 30 (TechNode)
Reference video clipsUp to 3 (ByteDance)Up to 10 (TechNode)
Reference audio clipsUp to 3 (ByteDance)Up to 10 (TechNode)
AudioTwo-channel stereo, made with the videoMade with the video; the API platform below lists it as an option you can switch off
Length setting on one API platform4 to 15 seconds4 to 30 seconds
Resolution options on one API platform480p, 720p, 1080p, 4K480p, 720p, 1080p
Aspect ratios on one API platform1:1, 4:3, 3:4, 16:9, 9:16, 21:9, adaptiveThe same list
Start from one imageYes, as a first frameYes, as a first frame

Two notes about this table. First, ByteDance's 2.5 page does not state a resolution, so we do not state one for the model itself. Second, on the API platform, a first-frame request and a reference request are separate modes and you cannot mix them in one request.

Seedance and other AI video models

Seedance is one of several strong video models. Kling from Kuaishou and Veo from Google are the two names people compare most often. Each has different limits for clip length, inputs and audio. We compare them in Seedance vs Kling vs Veo.

Seedance on Effect Farm

Effect Farm is an independent site at effectfarm.com. We are not part of ByteDance, and ByteDance does not endorse us. We make videos from your photo with ready-made templates, and those templates run on Seedance. We do not say which version here. The AI models page lists the models that make and check each video.

Each template page shows what you get:

You pay once per video. 30 credits (≈ 1 video) $4.99, 100 credits (≈ 3 videos) $12.90, 300 credits (≈ 10 videos) $29.90 and 1000 credits (≈ 33 videos) $79.90. Pack credits never expire. A failed video gives its credit back automatically.

FAQ

What is Seedance?

Seedance is ByteDance Seed's family of AI video models. It makes short videos with sound from text, images, audio or video. The versions in this guide are Seedance 2.0 and Seedance 2.5.

Who made Seedance?

ByteDance Seed made it. ByteDance Seed is a team inside ByteDance.

How do I use Seedance?

Use a ByteDance app, an API platform, or a ready-made template. For a dance from one photo, open a template page, upload a photo and pay for the video. You do not need to write a prompt.

Is Seedance free?

It depends on the platform. Each platform sets its own prices, limits and trial rules, and they change. Check the platform before you start. Effect Farm is paid: one video costs 30 credits ($4.99 in the smallest pack).

What is the difference between Seedance 2.0 and 2.5?

Seedance 2.5 makes longer clips (30 seconds against 15 seconds) and accepts more reference files. See Seedance 2.5 vs Seedance 2.0 for the full table.

Can Seedance make a video from one photo?

Yes. Seedance supports image-to-video. You can start from one image and describe the motion, or use a template that does it for you. See how to make an AI dance video from one photo.

Does Seedance make sound?

Yes. ByteDance describes Seedance as an audio-video model. Seedance 2.0 makes two-channel stereo audio with the video.

Does Effect Farm use Seedance?

Yes. The templates on effectfarm.com run on Seedance. We do not name the version on this page. See the AI models page for the models we use.

Try Seedance with one photo

Pick a template, upload one photo and get a video in 3–6 minutes. Start with Make Picture Dance or Genjutsu AI, or compare packs on the pricing page.

Sources

[1] ByteDance Seed, Seedance 1.0: https://seed.bytedance.com/en/seedance [2] ByteDance Seed, Seedance 1.5 Pro: https://seed.bytedance.com/en/seedance1_5_pro [3] ByteDance Seed, official launch of Seedance 2.0 (February 12, 2026): https://seed.bytedance.com/en/blog/official-launch-of-seedance-2-0 [4] ByteDance Seed, Seedance 2.5: https://seed.bytedance.com/en/seedance2_5 [5] TechNode, ByteDance launches Seedance 2.5 (July 31, 2026): https://technode.com/2026/07/31/bytedance-launches-seedance-2-5-video-generation-model/ [6] ByteDance Seed, Seedance 2.0 overview: https://seed.bytedance.com/en/seedance2_0 [7] API documentation for Seedance 2.0 (third-party platform): https://docs.kie.ai/market/bytedance/seedance-2.md [8] API documentation for Seedance 2.5 (third-party platform): https://docs.kie.ai/market/bytedance/seedance-2-5.md