AI Agent Hub
Back to skills
Seedance 2.0 Video Generation icon

Seedance 2.0 Video Generation

Design & Media Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please install @user_f95f21b8/dlazy-seedance-2-0 by following https://skillhub.cn/install/skillhub.md.

About this skill

Problem

seedance 2.0 wraps ByteDance's Seedance 2.0 video generation service through dLazy as a CLI entry point for agents or scripts. It addresses the operational details of calling an external hosted API: authentication, parameter passing, multimodal file upload, and asynchronous task polling, rather than local model deployment. For workflows that need short clips, video references, or first/last-frame transitions, it lets you pass prompts and local media paths without hand-writing HTTP requests.

How It Works

The core is a thin client: invoke dlazy seedance-2.0. Prompts and parameters go to api.dlazy.com; local image, video, or audio paths are uploaded to dLazy media storage at files.dlazy.com so the model can read them. Generated outputs are typically returned as URLs hosted on files.dlazy.com.

Supported capabilities include:
- Multimodal references: images, video, and audio can be used as input references.
- First/last-frame mode: frame inputs control the generated result.
- Text-to-video: text prompts generate video.

Parameters support pipe references such as -, @N, @N.path, @*, and @stdin, which helps chain upstream outputs. Async jobs can use --no-wait to return task.generateId and status, then poll with dlazy status <generateId> --wait.

Boundaries

Consider these constraints:
- Cloud SaaS: local files are uploaded, and outputs depend on dLazy hosting.
- Authentication and quota: a dLazy API key is required; insufficient_balance indicates low credit, while unauthorized indicates a missing or invalid key.
- Error handling: missing local files, missing parameters, safety-policy rejections, and other failures return distinct error codes, so agents should surface explicit guidance.

It is not suitable for offline generation, private inference, or workloads with strict data-residency requirements.

Use Cases

  • In ad production, use a reference video's style as a baseline, combined with text prompts, to generate a new product demo short video.
  • During video editing, use first/last-frame mode to generate smooth transition animation assets based on specified start and end frames.
  • In automated workflows, agents batch-call the CLI to submit text-to-video tasks and asynchronously poll for final video URLs.
  • In game motion design, input a static UI concept image and an audio track to generate a UI demo video with dynamic visual feedback.

Best For

  • Ad designers responsible for product demo videos, needing to convert reference videos and text prompts into standardized demo clips.
  • Backend engineers building automated video processing pipelines, needing to integrate CLI interfaces and handle async task polling and error codes.
  • Short video content creators, needing to quickly generate multimodal video assets based on first/last frames or audio/image references.
  • Engineers developing AI Agent prompts and toolchains, needing to call underlying video generation APIs and parse JSON-formatted output results.