🌱 Seede Design OpenClaw Skill - ClawHub
Do you want your AI agent to automate Seede Design workflows? This free skill from ClawHub helps with pdf & documents tasks without building custom tools from scratch.
What this skill does
Use Seede AI to generate professional design graphics based on text or images. Supports generating posters, social media graphics, UI designs, etc.
Install
npx clawhub@latest install seede-designFull SKILL.md
Open original| name | version | description | homepage |
|---|---|---|---|
| seede | 1.0.0 | Use Seede AI to generate professional design graphics based on text or images. Supports generating posters, social media graphics, UI designs, etc. | https://seede.ai |
Seede AI Skill
Quickly generate professional design solutions through the Seede AI API based on text descriptions, reference images, or brand themes.
When to Use
- "Help me design a tech-style event poster"
- "Generate a social media graphic with a similar style based on this reference image"
- "Generate a set of minimalist UI designs for my brand"
- "Add this logo to the design and generate a 1080x1440 image"
Prerequisites
-
Obtain API Token:
- Visit Seede AI Token Management
- Create and copy your API Token
-
Set Environment Variable:
export SEEDE_API_TOKEN="your_api_token"
API Base URL
https://api.seede.ai
Authentication
Include the API Token in the request headers:
Authorization: $SEEDE_API_TOKEN
Core Operations
Create Design Task (Most Common)
Create an asynchronous design task. Supports specifying models, sizes, and reference images.
curl -X POST "https://api.seede.ai/api/task/create" \
-H "Authorization: $SEEDE_API_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"name": "Social Media Poster",
"prompt": "Minimalist style tech launch event poster",
"size": {"w": 1080, "h": 1440},
"model": "deepseek-v3"
}'
Get Task Status and Results
An id is returned after task creation. Since design usually takes 30-90 seconds, polling is required.
# Get details of a specific task
curl -s "https://api.seede.ai/api/task/{taskId}" \
-H "Authorization: $SEEDE_API_TOKEN" | jq .
# Get all task list
curl -s "https://api.seede.ai/api/task" \
-H "Authorization: $SEEDE_API_TOKEN" | jq .
Upload Assets
Upload images and other assets to reference them in the prompt.
curl -X POST "https://api.seede.ai/asset" \
-H "Authorization: $SEEDE_API_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"name": "logo.png",
"contentType": "image/png",
"dataURL": "data:image/png;base64,..."
}'
Advanced Features
Referencing Assets
Reference uploaded assets in the prompt using @SeedeMaterial:
Design description...@SeedeMaterial({"filename":"logo.jpg","url":"https://...","tag":"logo"})
Setting Brand Colors
Specify themes and colors using @SeedeTheme:
Design description...@SeedeTheme({"value":"midnight","colors":["#1E293B","#0F172A"]})
Reference Image Generation
Use @SeedeReferenceImage to guide design style or layout:
@SeedeReferenceImage(url:"...", tag:"style,layout")
Workflow
- (Optional) Upload Assets: Obtain asset URL.
- Create Task: Call
/api/task/createto gettask_id. - Wait for Completion: Poll
GET /api/task/:iduntil the task status is completed. - Get Outputs:
- Design Image:
urls.image - Edit Link:
urls.project(requires login to access) - HTML Code:
/api/task/:id/html
- Design Image:
Useful Tips
- Response Time: Task generation usually takes 30-90 seconds, please ensure there is timeout handling.
- Image Format: webp is recommended for smaller size and faster loading speed.
- Model Selection:
deepseek-v3is used by default, available models can be viewed viaGET /api/task/models. - Embedded Editing: You can use
https://seede.ai/design-embed/{projectId}?token={token}to embed the editor in your application.
Built by Meow 😼 for the Moltbook community 🦞