Video Generate Skill for Openclaw

A professional-grade AI video generation skill leveraging Doubao Seedance models to transform text, images, and multimodal references into cinematic content.

warm-wm
v1.0.0
Mar 18, 2026
0
1.3k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install video-generate

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install video-generate using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Video Generate Skill?

The Video Generate skill is a robust integration designed for the Openclaw Skills ecosystem, enabling AI agents to produce high-quality video content using Doubao Seedance 1.0 and 1.5 models. It supports a wide range of creative workflows, from simple text-to-video prompts to complex multimodal transitions involving first-frame guidance, reference images, and synchronized audio.

This skill is engineered for versatility, offering granular control over cinematic parameters such as aspect ratio, resolution, and camera movement. By utilizing Openclaw Skills, developers can automate the production of marketing assets, social media content, or architectural visualizations with minimal manual intervention.

Video Generate Skill Use Cases

  • Creating cinematic 1080p videos from descriptive text prompts
  • Generating smooth transitions between two static images using first and last frame guidance
  • Applying specific visual styles to videos using up to four reference images
  • Producing multimodal content that combines video, image, and audio references
  • Synchronizing AI-generated audio with visual movement using Seedance 1.5 models

How Video Generate Skill Works

  1. The user provides a text prompt or reference material (images, video, or audio) to the AI agent.
  2. The skill validates the presence of necessary API keys like ARK_API_KEY within the environment.
  3. A generation request is sent to the Seedance model with specified parameters such as duration, ratio, and resolution.
  4. The system monitors the task status, performing polling or using a task ID for asynchronous queries.
  5. Upon completion, the skill saves the video to the local workspace and provides a direct URL and Markdown video embed for immediate viewing.

Video Generate Skill Setup

To use this within the Openclaw Skills framework, configure your environment variables first:

# Set your API Key (required)
export ARK_API_KEY="your_api_key_here"

# Optional: Set custom base URL or model name
export MODEL_VIDEO_API_BASE="https://api.example.com"
export MODEL_VIDEO_NAME="doubao-seedance-1-5-pro-251215"

You can then run generation tasks via the CLI:

python scripts/video_generate.py -p "A futuristic city" -n future_city -r 16:9 -d 5

Video Generate Skill Data Schema & Taxonomy

The skill manages outputs and metadata through a structured response format:

Field Description
video_name The unique identifier and filename for the output
success_list Contains mapping of video names to their accessible URLs
local_path Absolute path to the file in /root/.openclaw/workspace/skills/
task_id Unique ID used for polling or querying long-running tasks

Generated files are stored locally in the skill-specific workspace folder to ensure persistence across sessions.

Video Generate Skill Advanced Features

  • Multimodal synthesis combining 1-4 images, 0-3 videos, and 0-3 audio files as references
  • Intelligent model fallback logic to downgrade to stable versions if specific features are unavailable
  • Native support for Seedance 1.5 Pro audio generation capabilities
  • Customizable frame count logic following the $25 + 4n$ format for precise timing
  • Asynchronous task querying using unique task IDs to handle long generation windows without timeouts

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*