Jimeng, Dreamina & Seedance: ByteDance's AI Video, Explained

Jimeng, Dreamina, Seedance and Seedream, decoded: which is the app, which is the model, and how to use ByteDance's AI video outside China.

by HowAIWorks Team
On this page

Introduction

If you have seen the names Jimeng, Dreamina, Seedance and Seedream used interchangeably and assumed they must be one product, here is the one-sentence map: Jimeng and Dreamina are the same ByteDance creation app in two regions (Chinese and international), while Seedance is the video model and Seedream is the image model that both apps run on top of. In other words, two are apps and two are models, and they come from the same company. This post is the decoder for anyone trying to figure out whether Jimeng, Dreamina, Seedance, Seedream and CapCut are "all the same ByteDance thing." The short answer is that they are one family, and the confusion is mostly a branding artifact of ByteDance shipping the same technology under different names in different places.

The name map at a glance

ByteDance sells the same AI creation stack under several brands depending on where you are and which app you open. The cleanest way to hold it in your head is app versus model:

NameWhat it isWhere you meet it
Jimeng (即梦)The creation app, China-domesticChinese-language app, needs a +86 number or Douyin ID
DreaminaThe creation app, internationaldreamina.capcut.com, English interface
SeedanceThe video generation modelRuns inside Dreamina, Jimeng, CapCut, Doubao
SeedreamThe image generation modelRuns inside Dreamina, Jimeng, CapCut
CapCutByteDance's video editorSurfaces Seedance/Seedream as "Dreamina Seedance"
DoubaoByteDance's AI assistantAlso exposes some generation features

So when a leaderboard lists "Dreamina Seedance 2.0," it is naming the app and the model together — not describing a third product. The same generation engine appears in Doubao, ByteDance's consumer AI assistant, and inside CapCut, which is why the brand sprawl is easy to mistake for a family of separate tools.

Jimeng vs Dreamina: same app, two doors

The most common question — "is Dreamina the same as Jimeng?" — has a precise answer: they are the same product, regionalized. Jimeng (即梦) is the domestic Chinese app. It has a Chinese-language interface and, like most Chinese consumer apps, gates sign-up behind a mainland Chinese phone number (+86) or an existing Douyin account. That is the single biggest reason people outside China struggle to use it.

Dreamina is the international sibling. It lives at dreamina.capcut.com, presents an English interface, and lets you sign up with a Google account, an email address, or a TikTok login — no +86 number required. Functionally the two apps offer the same core: text-to-image, text-to-video, image-to-video, and avatar or talking-head generation, all powered by the same underlying models.

There are differences worth knowing. The domestic Jimeng app often gets newer model versions first and tends to be cheaper, while Dreamina occasionally trails on the very latest features but is far easier to reach. If you specifically want "Jimeng in English," the honest answer is that you do not translate Jimeng — you use Dreamina, which is the English-facing version of the same thing.

Seedance vs Seedream: the models under the hood

Underneath both apps sit two distinct ByteDance models, and mixing them up is the second most common confusion:

  • Seedance is the video model — text-to-video and image-to-video, with native audio in its latest generation. This is the one topping AI video leaderboards.
  • Seedream is the image model — text-to-image and image editing. As of mid-2026 the Dreamina interface exposes Seedream versions in the 4.x-to-5.0 range for still images.

A useful mnemonic: Seedance puts things in motion (video); Seedream paints a still picture (image). Both are members of ByteDance's broader "Seed" model family, and both are the kind of multimodal AI system that accepts text, images, and other inputs at once rather than a single modality.

Where Seedance stands

Seedance 2.0 launched in China on February 12, 2026. At launch it topped the Artificial Analysis text-to-video arena, ahead of Google Veo 3.1 and Kling 3.0, on the strength of a unified multimodal architecture with native audio-video synchronization — a genuinely strong result, though rankings shift depending on which board and category you look at. For the detailed capabilities, resolution and pricing breakdown, see our model page on Seedance 2.0.

The launch also triggered a Hollywood backlash — cease-and-desist letters from major studios over copyright — which delayed the global rollout. That story is its own piece; we cover it in full in Seedance 2.0: Breakthroughs and Copyright Launch Delay, so this post stays focused on the naming and access map.

A newer Seedance 2.5 was announced on June 23, 2026 at ByteDance's Volcano Engine conference, with the company targeting an early-July 2026 public launch that is still rolling out — Dreamina has begun surfacing it but still marks it "coming soon." ByteDance's own announcement centers on native 30-second single-shot generation, up to 50 multimodal reference inputs (up from 12), and native 4K output. Independent benchmark scores and pricing were not published at announcement, so treat any 2.5 performance claims as newer and less verified than the 2.0 numbers until ByteDance publishes a full model card.

How to use Seedance outside China

This is the practical intent behind most of the confusion, so here is the direct route. To use Seedance from outside China, you do not need a VPN, and you do not need a Chinese phone number. You have two clean options:

  1. Dreamina (dreamina.capcut.com). Open the site, sign in with Google, email, or TikTok, and select a Seedance version from the video model list. Free daily credits let you test before paying. This is the most beginner-friendly path.
  2. CapCut. ByteDance's editor exposes the same model — listed as Dreamina Seedance — inside its AI video tools, so a generated clip drops straight onto your editing timeline. Seedance rolled out through CapCut in stages across 2026, reaching the US, Europe, Japan and other markets.

The domestic Jimeng app is the one that requires a +86 number or a Douyin login, and virtual or roaming numbers are typically blocked, which is why "how to use Jimeng outside China" turns into a VPN-and-phone-number rabbit hole. Skip it: Dreamina is the same product without the barrier. Third-party API resellers also exist, but for most creators the official Dreamina or CapCut route is simpler and cheaper to start with.

How ByteDance's video stack compares

Seedance is ByteDance's entry in a crowded field of frontier video models. Its most-cited rivals are the other Chinese heavyweight, Kling from Kuaishou — available in the West through the Kling AI app — and MiniMax's Hailuo. On the Western side, the natural comparison is OpenAI's Sora, whose Sora 2 announcement set the bar for native audio-video generation that Seedance 2.0 was measured against at launch.

The practical takeaway for a Western creator: Seedance is reachable today through Dreamina, competitive at the top of the leaderboards, and generally cheaper per clip than the American frontier models — provided you go in through the international door rather than the domestic one.

Conclusion

The tangle of Jimeng, Dreamina, Seedance, Seedream and CapCut collapses into a simple structure once you separate apps from models. Jimeng and Dreamina are the same creation app in two regions; Seedance is the video model and Seedream is the image model that both run; CapCut and Doubao are additional ByteDance surfaces that call the same engines. If you are outside China and want to try ByteDance's AI video, the answer is almost always Dreamina — no VPN, no Chinese phone number, just the same Seedance model that tops the leaderboards, reached through the door that was built for you.

For more on accessing Chinese AI products from abroad, see what Quark is and how Cici became Dola.

Sources

Frequently Asked Questions

Almost. They are the same product line from ByteDance, split by region. Jimeng (即梦) is the China-domestic app with a Chinese interface; Dreamina is the international version at dreamina.capcut.com with an English interface. Both run the same underlying Seedance video model and Seedream image model.
They are two separate ByteDance models. Seedance is the video generation model (text-to-video and image-to-video). Seedream is the image generation model (text-to-image). Both are surfaced inside the Jimeng and Dreamina apps, so users often conflate the app name with the model names.
Use Dreamina at dreamina.capcut.com, or select the Seedance model inside CapCut. Neither requires a VPN or a Chinese phone number. The domestic Jimeng app is what needs a +86 number or a Douyin account, which is why it is hard to reach from outside China.
The domestic Jimeng app itself is Chinese-language. The English-facing product is Dreamina, ByteDance's international version of the same platform. If you want an English interface, use Dreamina rather than trying to translate Jimeng.
No. Dreamina is reachable globally and you sign up with Google, email, or a TikTok account, with no +86 number required. VPNs and Chinese phone numbers are only relevant to the domestic Jimeng app.
No. CapCut is ByteDance's video editor. Seedance is a generation model that CapCut can call, listed there as Dreamina Seedance. CapCut is one of several surfaces (alongside Dreamina, Jimeng and Doubao) that expose the same underlying model.

Continue Your AI Journey

Explore our lessons and glossary to deepen your understanding.