How to Use Seedance: The Complete Guide (2026)

ByteDance's Seedance has taken the AI video generation world by storm. Whether you're a content creator, developer, or just curious about the tool that's outpacing Sora 2 and Veo 3, this guide covers everything you need to know — from getting started for free to integrating the API into your workflow.

What Is Seedance?#

Seedance is an AI video generation model developed by ByteDance's Seed team. It generates high-quality video clips from text prompts, images, videos, and audio inputs. The model is known for exceptional character consistency, realistic physics, and native audio synchronization.

Seedance powers the video generation features inside ByteDance's consumer apps like Jimeng (即梦/Dreamina), Doubao (豆包), and Xiaoyunque (小云雀).

Seedance Versions Explained#

ByteDance has released several Seedance model variants. Here's how they compare:

Seedance 2.0 is the latest and most capable version, featuring:

• All-Round Reference System: Upload up to 12 files (9 images + 3 videos + 3 audio) and use @mentions to tell the AI exactly how to use each reference
• Native 2K Resolution: Generated natively, not upscaled
• Multi-Shot Storytelling: Describe multiple camera shots in a single prompt
• Native Audio & Lip-Sync: Generates synchronized dialogue, sound effects, and ambient audio
• Video Duration: 4–15 seconds per generation (precise to the second)
• Aspect Ratios: 16:9, 9:16, 4:3, 3:4, 21:9, 1:1

How to Access Seedance (All Methods)#

Method 1: Jimeng / Dreamina (Official Platform)#

Jimeng (即梦) is ByteDance's official creative platform and the primary home for Seedance.

China Version (jimeng.jianying.com):

• Full Seedance 2.0 access
• ~260 free credits for new users
• Paid plans start at ¥69/month (~$9.60 USD)
• ⚠️ Requires a Chinese phone number to register

International Version (dreamina.com):

• Seedance 1.0 and 1.5 available
• Seedance 2.0: Not yet available internationally (as of Feb 2026)
• Free tier: 60+ daily credits
• Paid plans: $18/month (Basic), $42/month (Standard), $84/month (Advanced)

Method 2: Doubao App (豆包)#

• 10 free video generations per day (5s or 10s clips)
• Requires joining the official Feishu (Lark) group and submitting your User ID for approval
• Approval takes 1–2 days
• ⚠️ Currently experiencing heavy congestion with long queue times

Method 3: Xiaoyunque App (小云雀) — Best Free Option#

• 1,200 bonus credits on registration
• 120 free credits daily (~12 seconds of video per day)
• 3 free generations on first login
• Subscription: ~¥39/month
• Currently in growth phase with generous free access
• Available on iOS App Store

Method 4: Third-Party Platforms (No Chinese Phone Number Needed)#

For users outside China who can't access the Chinese apps:

Method 5: Volcano Engine API (Official API)#

ByteDance's cloud platform offers official Seedance API access:

• Seedance 2.0 API launched February 14, 2026
• Enterprise-grade with SLA guarantees
• Requires Volcano Engine account
• Pay-per-use pricing

How to Register Without a Chinese Phone Number#

This is the #1 barrier for international users trying to access Seedance 2.0 on Jimeng, Doubao, or Xiaoyunque. Here are your options:

Option A: Use a Virtual Chinese Phone Number (Recommended for Quick Access)#

Several services provide temporary or virtual Chinese phone numbers for SMS verification:

1. SMS-Activate (sms-activate.guru)

   • One-time Chinese number for ~$0.50–$2.00
   • Select "China" → search for "ByteDance" or "Douyin" as the service
   • Receive the verification code and complete registration
   • Number is disposable — save your account credentials

2. 5SIM (5sim.net)

   • Similar service, Chinese numbers available
   • Pay-per-SMS model

3. MobileSMS.io

   • Offers Chinese numbers specifically for app verification

Step-by-step process:

1. Go to the virtual number service and purchase a Chinese number
2. Open the Jimeng/Doubao registration page
3. Enter the virtual Chinese number
4. Request the SMS verification code
5. Check the virtual number service for the incoming code
6. Enter the code to complete registration
7. Immediately bind an email address as backup login method

⚠️ Important: Virtual numbers may not always work if the platform detects them. If one number fails, try another. Prices are low enough that a few attempts are still cost-effective.

Option B: Use International Platforms Instead#

If you don't want to deal with Chinese phone numbers at all:

• Dreamina International (dreamina.com) — Sign up with Google/email, access Seedance 1.0/1.5
• Third-party wrappers like Pollo AI, Weshop AI — Access Seedance 2.0 through their interfaces
• API platforms like fal.ai, Replicate — Developer-friendly, no Chinese account needed

Option C: Buy a Pre-Registered Account#

Some sellers on platforms like Taobao or through social media offer pre-registered Jimeng accounts with credits. Prices range from $5–$20.

⚠️ Risk warning: Account sharing violates most platforms' ToS. Your account could be banned. Use at your own risk.

Seedance Prompt Guide: Tips & Techniques#

Basic Prompt Structure#

For Seedance 2.0, prompts work best when they're specific and cinematic:

[Subject] + [Action] + [Setting] + [Camera Movement] + [Mood/Lighting]

Example:

A young woman in a red dress walks through a rain-soaked Tokyo street at night.
Neon signs reflect off wet pavement. Camera slowly tracks behind her.
Warm orange and cool blue lighting contrast.

Using the @Reference System (Seedance 2.0)#

The @mention system is what makes Seedance 2.0 stand out. Upload reference files and tell the AI exactly how to use them:

Character Reference:

@Image1 is the main character. She walks into a coffee shop and sits down.
Close-up of her face as she smiles.

Motion Reference:

@Image1 as the character, perform the dance moves from @Video1.
Outdoor park setting, golden hour lighting.

Multi-Shot Storytelling:

Shot 1: Wide establishing shot of a mountain landscape at sunrise.
Shot 2: Medium shot of @Image1 standing at the cliff edge, wind blowing through hair.
Shot 3: Close-up of hands gripping a hiking pole.

Video Extension:

Extend @Video1 by 5s. The character continues walking and turns to look at the camera.

Combining Video + Audio:

@Image1 as cover, refer to @Video1's camera movement.
Background music follows the rhythm of @Audio1.

Pro Tips for Better Results#

1. Be cinematic, not technical — Describe scenes like a film director, not a programmer
2. Specify camera movements — "dolly in", "tracking shot", "slow pan left", "aerial view"
3. Include lighting details — "golden hour", "harsh overhead fluorescent", "candlelit"
4. Use high-quality references — 2K or 4K source images produce better results
5. Start with 5-second clips — Faster generation, iterate quickly before committing to longer durations
6. Adjust Creativity vs Consistency sliders — Start with 60% creativity / 40% consistency for balanced results
7. One concept per generation — Don't overload a single prompt with too many ideas
8. Iterate incrementally — Make small changes between generations rather than rewriting entire prompts

Prompt Examples by Use Case#

E-commerce Product Video:

A sleek wireless headphone sits on a marble surface. Camera slowly orbits 360 degrees.
Soft studio lighting with subtle reflections. Minimalist white background.

Music Video / Visualizer:

Abstract fluid shapes morph and pulse to the beat of @Audio1.
Vibrant neon colors against black background. Camera zooms in and out rhythmically.

Short Film / Drama:

@Image1 is a detective in a noir setting. He pushes open a heavy wooden door
and steps into a dimly lit bar. Smoke curls through the air.
Shot 1: Door opening from inside the bar. Shot 2: Close-up of his face, shadows across his eyes.

Social Media Content (9:16):

@Image1 looks directly at camera and speaks enthusiastically.
Bright, well-lit room with blurred background. Slight camera zoom in.
Natural, authentic vibe.

Seedance API Integration Guide#

fal.ai (Easiest for Developers)#

fal.ai offers the most developer-friendly access to Seedance 1.0 models:

Install the client:

npm install @fal-ai/client

Text-to-Video Example:

import { fal } from "@fal-ai/client";
 
const result = await fal.subscribe("fal-ai/seedance/text-to-video", {
  input: {
    prompt: "A cat playing piano in a jazz club, cinematic lighting",
    duration: 5,
    aspect_ratio: "16:9"
  }
});
 
console.log(result.data.video.url);

Image-to-Video Example:

const result = await fal.subscribe("fal-ai/seedance/image-to-video", {
  input: {
    prompt: "The character starts dancing gracefully",
    image_url: "https://example.com/character.jpg",
    duration: 5
  }
});

Available fal.ai Endpoints:

• fal-ai/seedance/text-to-video — Seedance 1.0 Pro text-to-video
• fal-ai/seedance/image-to-video — Seedance 1.0 Pro image-to-video
• fal-ai/seedance-lite — Seedance 1.0 Lite (faster, lower quality)

Replicate#

import replicate
 
output = replicate.run(
    "bytedance/seedance-v1-pro-i2v-480p",
    input={
        "prompt": "A woman walks through a field of sunflowers, golden hour",
        "image": "https://example.com/reference.jpg",
        "num_frames": 120,
        "fps": 24
    }
)
print(output)

Available Replicate Models:

• bytedance/seedance-v1-pro-i2v-480p
• bytedance/seedance-v1-pro-i2v-1080p
• bytedance/seedance-1-pro
• bytedance/seedance-1-pro-fast

Volcano Engine API (Official, Seedance 2.0)#

The official API through ByteDance's Volcano Engine supports the full Seedance 2.0 feature set:

Key Specs:

• Input: Up to 12 files (9 images, 3 videos, 3 audio) + text prompt
• Output: 4–15 second video, 480p–720p
• Image formats: JPEG, PNG, WebP, BMP, TIFF, GIF (max 30 MB each)
• Video formats: MP4, MOV (max 50 MB each)
• Audio: MP3 (max 15 MB each)

Access:

1. Create a Volcano Engine account at volcengine.com
2. Apply for Seedance API access
3. Generate API credentials
4. Use the REST API or official SDK

Note: The Volcano Engine API may require enterprise verification. For individual developers, fal.ai or Replicate are more accessible options for Seedance 1.0 models.

n8n / ComfyUI Integration#

For workflow automation:

n8n: Use the HTTP Request node to call fal.ai or Replicate APIs. Set up webhooks to trigger video generation from form submissions, social media posts, or scheduled events.

ComfyUI: Community nodes for Seedance are available on GitHub. Search for "ComfyUI-Seedance" to find integration nodes that connect to the fal.ai backend.

Seedance vs Competitors (2026)#

Frequently Asked Questions#

Is Seedance free?#

Yes, partially. Jimeng offers ~260 free credits for new users. Xiaoyunque gives 1,200 bonus credits + 120 daily. Doubao offers 10 free generations per day. Third-party platforms like HuggingFace have free demos with limited usage.

Is Seedance open source?#

No. Seedance is a proprietary model by ByteDance. However, Seedance 1.0 models are available through API platforms like fal.ai and Replicate.

Does Seedance generate audio?#

Seedance 1.5 Pro and 2.0 generate native audio including dialogue lip-sync, sound effects, and ambient sounds. Seedance 1.0 versions do not include audio.

Can I use Seedance for commercial projects?#

Yes, paid tier outputs on Dreamina are watermark-free and licensed for commercial use. Always check the specific platform's terms of service.

What's the maximum video length?#

Seedance 2.0 generates 4–15 second clips. You can extend videos by using the video extension feature to chain multiple generations together.

Do I need a Chinese phone number?#

For Jimeng (China), Doubao, and Xiaoyunque — yes, a Chinese phone number is required for registration. See our registration guide above for workarounds. For Dreamina International, fal.ai, Replicate, and other third-party platforms — no Chinese number needed.

Last updated: February 13, 2026

This guide is independently maintained and not affiliated with ByteDance.