AI Agent Hub
Back to skills
Doubao Seedream 4.5 Video Generation icon

Doubao Seedream 4.5 Video Generation

Design & Media Updated 2026.08.29

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

Follow https://skillhub.cn/install/skillhub.md to install @user_f95f21b8/dlazy-seedream-4-5.

About this skill

Problem Being Solved

When you want to bring Doubao Seedream 4.5 into an agent, script, or content workflow, you usually need to handle API authentication, video parameters, local asset upload, async task status, and hosted output URLs. This skill wraps that flow around a dlazy seedream-4.5 command, making it easier to call from the command line or agent tooling.

How It Works

The skill is a thin client over the dLazy hosted API. Its core capabilities are text-to-video and image-to-video generation. It requires a dLazy API key, which can also be provided per invocation through the DLAZY_API_KEY environment variable. If an image or video field receives a local file path, the skill uploads the file to dLazy media storage before the model reads it, and the generated output is returned as a hosted URL.

  • Use flags such as --prompt to describe the target video
  • Submit async work with --no-wait, then poll with dlazy status --wait
  • Results can include output URLs; async responses include task.generateId and status
  • Recovery depends on error codes such as unauthorized, insufficient_balance, missing required parameters, or local file read failures

Boundaries And Caveats

It depends on a dLazy account, organization scope, and balance, so it is not a local inference tool. Local image or video files are uploaded to files.dlazy.com, and generated outputs are also hosted in the cloud. It fits engineers who want Seedream 4.5 in automation; if media must stay on-premises or the workflow requires a fully offline video generation pipeline, this skill is not a suitable choice.

Use Cases

  • Use an agent prompt to call Seedream 4.5 for short video clips.
  • Pass an existing product image to generate a short demo clip.
  • Submit text-to-video jobs in a script, poll status, collect URLs.
  • Map error codes to missing key, low balance, or bad local file.

Best For

  • Backend engineers adding video generation to an agent toolchain.
  • Content designers turning static product images into demo clips.
  • Platform engineers scripting batch clips and tracking async outputs.
  • Application developers integrating dLazy auth and balance errors.