Introducing BytePlus Models “Seedance, Seedream, and Seed” on Puter.js in Collaboration With the BytePlus Team
On this page
We're excited to announce our partnership with BytePlus, ByteDance's model platform, bringing the Seed, Seedream, and Seedance families to Puter.js and our community of over 80,000 developers. Because the integration is direct, new models land in Puter.js as they launch, and developers can build with ByteDance's chat, image, and video models with no API keys, no accounts, and no server-side setup.
About BytePlus
BytePlus is ByteDance's enterprise cloud arm, an AI-native cloud running end to end from training to serving. Its Model Ark platform is the model layer, an all-in-one platform that serves ByteDance's own frontier models alongside models from other labs, behind low-latency, high-concurrency APIs.
The models ByteDance builds itself come from its Seed team, which covers three distinct families.
Seed is the language and multimodal line. Its models take text, image, and video input and return text, with context windows up to 256K tokens, tool calling, and adjustable reasoning depth. The tiers run from small high-throughput models to flagship reasoning and coding models.
Seedream handles images, with generation and editing in a single model. It takes multiple reference images to keep characters and styles consistent across a series, outputs up to 4K, and renders text in a range of languages.
Seedance handles video, generating clips from text prompts or reference images at 480p through 4K and 24fps, with subjects and style held steady across multiple shots. The newer models produce audio in the same pass as the video, with lip sync.
Models Available in Puter
Chat and Multimodal
- Dola Seed 2.1 Turbo: Faster, lower-cost Seed 2.1 variant with strong video understanding scores
- Seed 2.0 Lite and Seed 2.0 Mini: Mid-tier and small tiers for high-volume production workloads
- Seed 2.0 Pro: Top tier of the Seed 2.0 family for reasoning and coding, 256K context, text plus image and video input
- Seed 2.0 Code Preview: Coding-focused variant for repository-level agentic work
Image
- Dola Seedream 5.0 Pro: Top Seedream tier, with text rendering across 14+ languages
- Seedream 5.0 Lite: Cost-optimized tier that runs a reasoning pass before generating
- Seedream 4.5: Editing-focused update with 4K output at no extra charge
- Seedream 4.0: Unified generation and editing, up to 10 reference images
Video
- Dreamina Seedance 2.0 Mini: Cheapest tier of the 2.0 family, capped at 720p
- Dreamina Seedance 2.0: Top tier, up to 4K and 15 seconds, accepts text, image, video, and audio references
- Dreamina Seedance 2.0 Fast: Same motion engine at 720p and 480p for faster turnaround
- Seedance 1.5 Pro: Joint audio and video generation with multi-language lip sync
- Seedance 1.0 Pro: Multi-shot generation with consistent subjects across cuts
Models from Other Labs
Alongside its own models, BytePlus hosts popular open-weight models from other labs.
- GLM-5.2: Z.ai's open-weight MoE model for agentic coding, with a 1M token context window
- DeepSeek V4 Pro and DeepSeek V4 Flash: 1M context MoE models for long-running coding agents
- GPT-OSS 120B: OpenAI's Apache 2.0 reasoning model with adjustable reasoning effort
Get Started
All of these models are ready to use right now through Puter.js. Add one script tag to your HTML:
<script src="https://js.puter.com/v2/"></script>
Or install via npm:
// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';
Then call any BytePlus model with a few lines of code:
<html>
<body>
<script src="https://js.puter.com/v2/"></script>
<script>
// Chat with Seed 2.0 Pro
puter.ai.chat(`Hello, how are you?`, {
model: 'byteplus/seed-2-0-pro-260328',
}).then(puter.print);
</script>
</body>
</html>
Image generation works the same way and resolves to an HTMLImageElement:
puter.ai.txt2img(
"A layered product shot of a ceramic mug on a linen backdrop",
{ model: "byteplus/dola-seedream-5-0-pro-260628" }
).then(image => {
document.body.appendChild(image);
});
And video:
puter.ai.txt2vid(
"A drone shot sweeping over a misty mountain valley at sunrise",
{ model: "byteplus/dreamina-seedance-2-0-260128" }
).then(video => {
document.body.appendChild(video);
video.play();
});
That's it. No API keys and no infrastructure setup.
Looking Ahead
BytePlus covers text, vision, and speech on Model Ark, served at low latency and high concurrency. Its chat, image, and video models now run through Puter.js in a single integration, so all three modalities come from the same library.
We'd like to thank the BytePlus team for their collaboration and support in making this integration possible. We're glad to partner with a team that shares our goal of making AI accessible to every developer.
This is just the beginning. Stay tuned for more updates as we continue to expand what's possible with Puter and BytePlus.
Related
Ship a Full-Stack App with One Prompt
Build an AI chat app using Puter.js
Coding manually? see the guide