A high-performance image generation and editing skill powered by ByteDance's Seedream models for creating professional visuals, posters, and brand assets.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install seedream
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install seedream 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 Seedream?
Seedream is a sophisticated image generation suite developed by ByteDance, now accessible as one of the premier Openclaw Skills. It specializes in high-fidelity text-to-image creation, complex image editing, and batch sequential generation. With a focus on typography excellence and poster design, it allows developers and creators to generate images at up to 4K resolution with built-in prompt optimization.
The skill leverages the Atlas Cloud API to provide access to multiple model generations, including the flagship v5.0 Lite. This version introduces PNG output support and rapid generation speeds of 2-3 seconds, making it an ideal choice for integrated AI workflows that require both speed and aesthetic precision.
Seedream Use Cases
- Creating marketing visuals and social media graphics with precise typography and layout control.
- Designing professional posters, movie art, or brand assets using specialized design-centric models.
- Batch generating a series of related product photos or storyboard frames in a single request.
- Editing existing photographs to change backgrounds or styles while preserving subjects and lighting.
- Generating high-resolution 4K concept art for games, websites, or digital media.
How Seedream Works
- The user provides a descriptive text prompt or an image URL for modification.
- The skill selects the appropriate Seedream model variant (v5.0 Lite, v4.5, or v4) based on the task requirements.
- A POST request is sent to the Atlas Cloud API to initiate the generation or editing task.
- The system enters a polling loop, checking the prediction status every 3 seconds until the generation is complete.
- Once the status reaches success, the final image URLs or Base64 data are retrieved.
- The resulting assets are downloaded locally or presented to the user for further use.
Seedream Setup
To get started with this among your Openclaw Skills, follow these steps:
- Sign up for an account at Atlas Cloud.
- Navigate to the Console -> API Keys and create a new key.
- Configure your environment variable:
export ATLASCLOUD_API_KEY="your-key-here"
- You can now use the included Python script to list models or generate images:
python scripts/generate_image.py generate --model "bytedance/seedream-v5.0-lite" --prompt "A futuristic city" --output ./output
Seedream Data Schema & Taxonomy
The skill organizes its operations through the Atlas Cloud API, utilizing the following structure:
| Data Type | Description | Key Fields |
|---|---|---|
| Request | Generation Parameters | model, prompt, size, output_format |
| Prediction | Task Metadata | id, status, outputs (Array of URLs) |
| Configuration | Model Variants | v5.0 Lite, v4.5, v4 (Text, Edit, Sequential) |
All generated files are typically saved to a local ./output directory in JPEG or PNG format.
Seedream Advanced Features
- Batch Sequential Generation: Create up to 15 related images or edited variations in a single API call to maintain stylistic consistency.
- Prompt Optimization: Toggle between Standard and Fast modes to automatically enhance user prompts for better aesthetic results.
- Multi-reference Editing: Support for up to 14 reference images to guide complex style transfers and visual modifications.
- Typography Excellence: Enhanced engine for rendering legible, stylized text within generated posters and graphics.
- 4K Tier Support: Access to 16 preset resolutions including ultrawide and high-density formats like 4704x2016.
SKILL.md
Loading
Related Openclaw Skills
Submit PR
A standardized Pull Request workflow that automates security scanning, file verification, and conventional commit generation.
8004: ERC-8004 Agent Trust Protocol
A decentralized trust infrastructure for AI agents to discover, identify, and verify each other across organizational boundaries on the Celo network.
Memorine
A pure Python and SQLite persistent memory layer that gives AI agents human-like capabilities like fact retention, event causal chains, and automatic forgetting.
Huisuiyun Invoice Verification
A specialized tool for verifying the authenticity and details of Chinese invoices using the official Huisuiyun API.
Restart Task Recovery
A robust workflow designed to preserve context and resume in-progress multi-agent tasks across system restarts or configuration changes.
RDK X5 GPIO
A comprehensive hardware interface skill for controlling the RDK X5 40-pin header via GPIO, PWM, and standard serial protocols.
Featured*
