
How to Access the Seedance 2.0 API: Complete Guide for International Developers (2026)

ByteDance officially released Seedance 2.0 on February 12, 2026. The API opens in late February. Here's exactly how to get access — whether you're an international developer, a Chinese enterprise, or just someone who wants to test it before committing.
TL;DR — Seedance 2.0 API Access Options
| Access Method | Available Now? | Requires Chinese Phone? | Best For |
|---|---|---|---|
| Third-party API (EvoLink, kie.ai, Atlas Cloud) | ⏳ Full support late February | No | International developers, production use |
| Volcengine / BytePlus (Official) | ⏳ Workstation only; API late February | Yes (or enterprise verification) | Chinese developers, enterprise |
| Jimeng Web UI (即梦) | ✅ Live now | Yes (Douyin account) | Testing, prototyping, non-developers |
What Is the Seedance 2.0 API?
@image1, @video2, @audio3 syntax. The model uses these references to control character appearance, camera movement, choreography, and audio sync.
Why you need API access (not just the web UI):
- Batch processing — generate hundreds of product videos programmatically
- Pipeline integration — chain Seedance 2.0 into your editing or marketing workflows
- Custom control — set exact resolution, duration, aspect ratio, and reference files per request
- Cost efficiency — API calls are cheaper per generation than web UI credits at scale
Key upgrades over Seedance 1.5:
| Feature | Seedance 1.5 Pro | Seedance 2.0 |
|---|---|---|
| Max resolution | 1080p | 2K |
| Native audio | ✅ | ✅ Stereo, lip-sync in 8+ languages |
| @ reference system | ❌ | ✅ Up to 12 files |
| Multi-shot storytelling | ❌ | ✅ Coherent multi-scene videos |
| Video editing capabilities | Basic | Advanced (scene transitions, cuts) |
| Video duration | 5–15 sec | 4–15 sec |
Seedance 2.0 API Release Date and Current Status
What's available RIGHT NOW (as of February 17, 2026):
- ✅ Jimeng web interface — full Seedance 2.0 access (requires Chinese Douyin account)
- ✅ Xiao Yunque app — free tier with 3 generations + daily credits
- ✅ Doubao app — Seedance 2.0 integrated into ByteDance's AI assistant
- ✅ Volcengine workstation — enterprise testing interface (no programmatic API yet)
- ✅ Seedance 1.5 Pro API — available now via Volcengine for developers who need something today
- ⏳ Third-party platforms — full Seedance 2.0 support confirmed for late February
What launches in late February:
- Seedance 2.0 REST API via Volcengine / BytePlus
- Third-party platform integrations (EvoLink, kie.ai, Atlas Cloud, and others)
- Full async job-based API with text-to-video and image-to-video endpoints
Method 1 — Access via Third-Party API Platforms (Recommended for International Developers)
If you don't have a Chinese phone number — or you just want a simpler integration experience — third-party API platforms are the way to go.
Why third-party platforms make sense:
- No Chinese phone number or enterprise verification required
- Unified API — same endpoint and SDK for Seedance 2.0, Kling, Sora 2, Veo 3, and other models
- OpenAI-compatible format — minimal code changes if you're already using OpenAI's SDK
- Often cheaper — smart routing and volume discounts can reduce costs by 20–70%
EvoLink — Step-by-Step
- Sign up at evolink.ai/signup — takes 30 seconds, free credits included, no credit card required
- Create an API key in the dashboard — one-click generation
- Update your base URL to EvoLink's endpoint and set the model to
seedance-2.0 - Make your first API call
Other Third-Party Options
- kie.ai — another API aggregator with video model support; check their pricing page for Seedance 2.0 availability
- Atlas Cloud — offers a $1 registration bonus; has announced day-one Seedance 2.0 integration
- Apiyi — China-focused aggregator with OpenAI-compatible format; claims ~10% discount vs. official pricing
Pick whichever platform fits your existing stack. The API patterns are nearly identical across all of them.
Method 2 — Access via Volcengine (Official, China-Based)
This is the direct route from ByteDance's cloud platform. It's the right choice if you're a Chinese developer, need enterprise SLAs, or want guaranteed first-party support.
Requirements:
- Chinese phone number or enterprise business verification
- Volcengine account (volcengine.com)
- For international users: BytePlus account (byteplus.com)
Steps:
- Register on Volcengine (or BytePlus for international)
- Navigate to the Volcano Ark (火山方舟) console
- Apply for Seedance 2.0 API access (available from late February)
- Generate API credentials (API key + secret)
- Integrate using their SDK or REST endpoints
Limitations:
- Region-restricted — some features may be limited outside mainland China
- Enterprise-focused — individual developers may find the onboarding heavier than third-party alternatives
- Separate accounts per service — unlike unified platforms, you'll need a distinct Volcengine integration
Method 3 — Use Seedance 2.0 via Web Interface (No API)
Jimeng (即梦) — jimeng.jianying.com
This is ByteDance's flagship AI creation platform and the most complete way to use Seedance 2.0 today. You get full access to all features: text-to-video, image-to-video, the @ reference system, multi-shot storytelling, and native audio generation.
- Visit jimeng.jianying.com
- Log in with a Douyin account
- Go to Generate → Video Generation → Select Seedance 2.0
- Enter your prompt or upload reference materials
- Dreamina — the international-facing version of Jimeng (more limited feature set)
- Xiao Yunque (小云雀) — free tier with 3 generations + 120 daily points
Important: This is great for testing, NOT for production. The web interface has no batch processing, hard rate limits, and zero programmatic access. If you need more than a few videos per day, you need the API.
Seedance 2.0 API Pricing: What to Expect
Pricing varies by platform; check the Volcengine console and docs for the latest token-based pricing (as of February 17, 2026, Volcengine shows per-token rates for voiced/unvoiced video). Here's what we know from third-party platforms.
Estimated official pricing (based on third-party source estimates and Seedance 1.5 Pro as baseline):
| Tier | Resolution | Estimated Price |
|---|---|---|
| Basic | 720p | ~$0.10/min of video |
| Standard | 1080p | ~$0.30–$0.50/min |
| Cinema | 2K | ~$0.60–$0.80/min |
Third-party pricing (what we know so far):
- EvoLink — launching with discount pricing in late February; exact rates TBA, but their model is to undercut direct provider pricing by 20–70% through smart routing
- kie.ai — has a pricing page for video models; check for Seedance 2.0 listing after late February
- Atlas Cloud — $1 registration bonus for new users; per-generation pricing TBA
- Apiyi — claims ~10% discount vs. official Volcengine pricing
For context: Competing models charge significantly more. Sora 2 runs approximately $0.50–$1.00 per generation for similar quality, and Runway Gen-4 uses a per-second credit model ($0.01/credit, with models consuming 5-12 credits/second depending on tier). Seedance 2.0 is expected to be among the most cost-effective options on the market.
Check back after late February for confirmed numbers. We'll update this section once official and third-party pricing is live.
Troubleshooting Common Issues
"Access denied" or region restrictions
"Model not available" error
- Verify the date — if it's before late February, the model endpoint may not be live yet
- Double-check the model string:
seedance-2.0(notseedance-2,seedance2.0, etc.) - If using Volcengine directly, confirm your account has been approved for Seedance 2.0 access
Rate limits and throttling
- Implement a job queue with concurrency limits (e.g.,
asyncio.Semaphorein Python) - Use exponential backoff on 429 responses
- For high-volume workloads (100+ generations/day), contact your provider about enterprise tier limits
- Batch your prompts and spread submissions over time windows
Face censorship or content filtering
- Use stylized or semi-realistic character prompts instead of photorealistic faces
- Avoid prompts referencing real people by name
- For product videos: focus on objects, environments, and abstract characters
Audio sync issues
- Keep videos under 10 seconds for best audio-visual sync
- Provide reference audio via
@audio1rather than relying on auto-generated audio - Specify audio cues in your prompt: "upbeat music starting at the beginning, sound effect of footsteps"
Seedance 2.0 API vs. Competitors — Quick Comparison
| Seedance 2.0 | Sora 2 | Kling 3.0 | |
|---|---|---|---|
| Developer | ByteDance | OpenAI | Kuaishou |
| Max resolution | 2K | 1792×1024 | 1080p |
| Native audio | ✅ Stereo, lip-sync | ✅ Full audio (dialogue, SFX, BGM) | ✅ Multilingual |
| Multimodal input | 9 img + 3 vid + 3 audio | Text + image | Text + image |
| Multi-shot | ✅ | ❌ | ❌ |
| Video duration | 4–15 sec | Up to 12s (API: 4/8/12s) | 3–15 sec |
| Est. price/min | $0.10–$0.80 | $0.50–$1.00 | $0.20–$0.60 |
| API available | Late February 2026 | Live | Live |
Seedance 2.0's killer advantages are the @ reference system and native stereo audio. Sora 2 leads in physics accuracy and photorealism. Kling 3.0 wins on motion naturalness and cost-effectiveness. Choose based on your specific needs.
FAQ
When does the Seedance 2.0 API launch?
Can I access the Seedance 2.0 API from outside China?
Is Seedance 2.0 open source? Can I run it locally?
No. Seedance 2.0 is a closed, API-only model. ByteDance has not released model weights, and there's no indication they plan to. You cannot run Seedance 2.0 locally — all inference happens on ByteDance's (or a third-party provider's) infrastructure. While ByteDance has open-sourced some language models, Seedance 2.0 video generation remains proprietary.
How much does the Seedance 2.0 API cost?
Pricing varies by platform. Volcengine console shows token-based pricing for voiced/unvoiced video - check their latest docs for exact rates. Third-party platforms may offer discounts of 10-70%. EvoLink is launching with a promotional discount in late February.
What's the difference between Seedance 2.0 and Seedance 1.5?
The major upgrades: 2K resolution (up from 1080p), native stereo audio generation with lip-sync in 8+ languages, the @ reference system supporting up to 12 input files (images, videos, audio), multi-shot storytelling for coherent multi-scene narratives, and improved video editing capabilities. The underlying architecture shifted to a Dual-Branch Diffusion Transformer, enabling audio-video joint generation in a single pass.
Can I use Seedance 2.0 for commercial projects?
Yes. API access includes commercial usage rights. Whether you're generating product videos, marketing content, or creative assets for clients, Seedance 2.0 output can be used commercially. Review ByteDance's specific terms of service (available on Volcengine) or your third-party provider's license terms for details on attribution requirements and content restrictions.
Which third-party platform is cheapest for Seedance 2.0?
It's too early to give a definitive answer — most platforms haven't published final Seedance 2.0 pricing yet. EvoLink is launching with a promotional discount in late February. Atlas Cloud is offering a $1 registration bonus. After late February, compare actual per-generation costs across platforms. The best deal will depend on your volume: high-volume users should look at tiered or enterprise pricing, while low-volume users may benefit from platforms with free credits.


