AI image generation with GPT Image 2, built for OpenClaw, Claude Code, OpenCode, and AI agents that need fast OpenAI-compatible image generation workflows.
Menu • Install • Showcase • GPT Image 2 • API Key • EvoLink
AI Agent? Skip the README, go straight to llms-install.md for step-by-step installation instructions designed for you.
- What is This?
- Installation
- Getting an API Key
- Showcase
- GPT Image 2 Generation
- File Structure
- Troubleshooting
- Compatibility
- License
- Community
GPT Image 2 Gen Skill is an AI agent image generation skill for OpenClaw, Claude Code, and OpenCode, powered by EvoLink. It gives your agent a fast OpenAI-compatible API integration for GPT Image 2, including text-to-image, image editing, batch generation, and multi-size output workflows.
| Skill | Description | Model |
|---|---|---|
| GPT Image 2 Gen | Text-to-image, image editing, batch generation | GPT Image 2 (OpenAI) |
openclaw skills add https://github.com/EvoLinkAI/gpt-image-2-gen-skillnpx evolink-gpt-imageOr non-interactive (for AI agents / CI):
npx evolink-gpt-image -yInstall to a specific directory:
npx evolink-gpt-image -y --path ~/.claude/skillsgit clone https://github.com/EvoLinkAI/gpt-image-2-gen-skill.git
cd gpt-image-2-gen-skill
openclaw skills add .Tell your AI agent the following prompt, and it will install the skill automatically:
Install the GPT Image 2 generation skill by running:
npx evolink-gpt-image@latest -y --path ~/.claude/skills
After installation, set the API key:
export EVOLINK_API_KEY=your_key_here
Then read the skill file at ~/.claude/skills/gpt-image-2-gen/SKILL.md to learn how to use it.
Install the GPT Image 2 generation skill by running:
npx evolink-gpt-image@latest -y --path ~/.opencode/skills
After installation, set the API key:
export EVOLINK_API_KEY=your_key_here
Then read the skill file at ~/.opencode/skills/gpt-image-2-gen/SKILL.md to learn how to use it.
Install the GPT Image 2 generation skill by running:
npx evolink-gpt-image@latest -y
The installer will auto-detect your OpenClaw skills directory. After installation, set the API key:
export EVOLINK_API_KEY=your_key_here
For agents that support shell commands, this single command installs and verifies in one step:
EVOLINK_API_KEY=your_key_here npx evolink-gpt-image@latest -y --path ~/.claude/skillsReplace ~/.claude/skills with ~/.opencode/skills or your agent's skill directory.
- Sign up at evolink.ai
- Go to Dashboard -> API Keys
- Create a new key
- Set it in your environment:
export EVOLINK_API_KEY=your_key_hereOr tell your AI agent: "Set my EvoLink API key to ..." — it will handle the rest.
| Portrait Styling | Product Marketing | Character Design |
|---|---|---|
 |
 |
 |
These examples were selected from the awesome-gpt-image-2-prompts repository. Install the skill, connect your EvoLink API key, and use the GPT Image 2 image generation docs to create similar outputs.
Generate and edit AI images through natural conversation with your AI agent.
- Text-to-image — Describe what you want, get an image
- Image editing — Provide reference images (1-16) and describe edits
- Batch generation — Generate up to 10 images per request
- Multiple sizes — 15 ratio presets + custom pixel dimensions
- Resolution tiers — 1K (~1MP), 2K (~4MP), 4K (~8.3MP)
- Quality levels — Low (fast), Medium (balanced), High (best)
- Prompt power — Up to 32,000 characters per prompt
Just talk to your agent:
"Generate an image of a sunset over the ocean"
"Create a minimalist logo, 1024x1024, high quality"
"Edit this image — add a cat next to the person"
"Generate 4 variations of a pixel art robot in 4K"
The agent will guide you through any missing details and handle the generation.
curlandjqinstalled on your systemEVOLINK_API_KEYenvironment variable set
The skill includes scripts/gpt-image-gen.sh for direct command-line use:
# Text-to-image (basic)
./scripts/gpt-image-gen.sh "A beautiful sunset over the ocean"
# High quality 4K widescreen
./scripts/gpt-image-gen.sh "Cinematic cityscape at dusk" --size 16:9 --resolution 4K --quality high
# Custom pixel dimensions
./scripts/gpt-image-gen.sh "Minimalist logo" --size 1024x1024
# Image editing
./scripts/gpt-image-gen.sh "Add a cat next to her" --image "https://example.com/photo.png"
# Batch generation
./scripts/gpt-image-gen.sh "Pixel art robot" --count 4 --quality high
# Dry run (preview payload)
./scripts/gpt-image-gen.sh "Test prompt" --dry-runSee references/api-params.md for complete API documentation.
.
├── README.md # This file
├── SKILL.md # Skill definition (for AI agents)
├── _meta.json # Skill metadata
├── bin/
│ └── cli.js # npm installer CLI
├── references/
│ └── api-params.md # Complete API parameter reference
└── scripts/
└── gpt-image-gen.sh # Image generation script
| Issue | Solution |
|---|---|
jq: command not found |
Install jq: apt install jq / brew install jq |
401 Unauthorized |
Check your EVOLINK_API_KEY at evolink.ai/dashboard |
402 Payment Required |
Add credits at evolink.ai/dashboard |
Content blocked |
Prompt flagged by moderation — modify your description |
| Image too large | Reference images must be <=50MB each |
| Generation timeout | Images can take 5-90s. Try lower quality/resolution first. |
| Agent | Install Method |
|---|---|
| OpenClaw | openclaw skills add <repo> or npx evolink-gpt-image |
| Claude Code | npx evolink-gpt-image -y --path ~/.claude/skills |
| OpenCode | npx evolink-gpt-image -y --path ~/.opencode/skills |
| Cursor | npx evolink-gpt-image -y --path <your-skills-dir> |
MIT
- Docs: GPT Image 2 image generation manual
- EvoLink: Create an API key
- Workflow repo: GPT-Image-2 × Seedance 2.0
- Follow EvoLinkAI: x.com/EvoLinkAI
Powered by EvoLink , Unified AI API Gateway