PickSkill
← Back

Seede Design

Use Seede AI to generate professional design graphics based on text or images. Supports generating posters, social media graphics, UI designs, etc.

README.md
Rendered from GitHub raw
View raw ↗

Seede Skill for Open Claw

The Ultimate AI Design Skill for Open Claw & Agents. Generate professional UI, social media graphics, posters, and more directly via Open Claw.

This repository contains the Seede Skill, enabling Open Claw (and other compatible agents) to generate high-quality designs using Seede AI.

Why Seede Skill?

  • 🚀 State-of-the-Art Generation: Uses the latest AI models to create high-quality, editable designs.
  • 🤖 Agent-First: Designed specifically for autonomous agents to control design parameters.
  • 🎨 Brand Consistency: Supports brand colors and asset injection.
  • 🛠️ Full Control: Precise control over resolution, format, and scene types.

Features

  • Text to Design: Generate complex designs from natural language.
  • Asset Management: Upload and manage logos, product shots, and reference images.
  • Design Management: List, search, and retrieve design URLs.
  • Token Management: Create and manage API tokens directly from the CLI.

Installation

To add this skill to your Open Claw agent, simply run:

npx skills add seedeai/seede-skill

Standalone CLI Installation

If you wish to use the tool manually without an agent:

npm install -g seede-cli

Configuration

The skill requires the SEEDE_API_TOKEN environment variable to be set to authenticate with the Seede AI service.

1. Create a Token:

You can generate a token directly via the CLI:

seede token create --name "My Agent Token"

2. Set Environment Variable:

export SEEDE_API_TOKEN="your_generated_token"

Usage

As an Open Claw Skill

Once installed, you can ask Open Claw to perform design tasks using natural language:

"Design a modern tech conference poster with neon accents." "Create a social media banner for a coffee shop."

The agent will automatically utilize the seede-design skill to fulfill these requests.

Manual Usage (CLI)

You can also use the CLI directly for testing or manual generation:

Interactive Mode:

seede create

Command Line (Agent Mode):

seede create --no-interactive \
  --prompt "Modern tech conference poster with neon accents" \
  --scene "poster" \
  --format "png"

Advanced Usage

Integrating User Assets

To place a specific image (like a logo or product shot) into the design:

  1. Upload the file first using seede upload.
  2. Reference the returned URL in the prompt using @SeedeMaterial:
seede create --no-interactive \
  --prompt "Minimalist product poster featuring this item @SeedeMaterial({'url':'<ASSET_URL>','tag':'product'})" \
  --scene "poster"

Enforcing Brand Guidelines

To ensure the design matches specific brand colors:

seede create --no-interactive \
  --prompt "Corporate annual report cover @SeedeTheme({'colors':['#000000','#FFD700']})"

Manage API Tokens

You can create and manage API tokens for CI/CD or Agent integration directly from the CLI.

Create a Token:

seede token create --name "My Agent Token" --expiration 30

Skill Metadata

This repository includes a SKILL.md file that defines the skill capabilities and metadata for the agent system.