A comprehensive local image generation tool for Openclaw Skills that leverages Fooocus and Stable Diffusion XL for offline visual creation.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install fooocus-image-gen
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 fooocus-image-gen 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 Fooocus Image Generation Skill?
The Fooocus Image Generation Skill integrates the powerful, open-source Fooocus software directly into your AI agent environment. It is designed for users who need to generate high-quality images locally using Stable Diffusion XL (SDXL) without the privacy concerns or costs associated with cloud-based APIs. This skill provides a streamlined interface for complex tasks such as text-to-image generation, image variations, and face swapping while maintaining a professional and developer-friendly workflow for Openclaw Skills users.
By leveraging Fooocus, the skill hides the inherent complexity of SDXL behind a user-friendly logic while exposing advanced features like inpainting and outpainting. It is particularly valuable for offline environments or projects requiring high keyword density and specific artistic control through local models and LoRAs.
Fooocus Image Generation Skill Use Cases
- Creating high-fidelity character portraits and scenery using local SDXL models.
- Transforming existing images through style transfer, upscaling, or artistic variations.
- Editing specific regions of an image using manual inpainting or extending canvas boundaries via outpainting.
- Generating consistent 2D anime or photorealistic content using specialized startup presets.
- Extracting prompts and metadata from existing images to learn and replicate specific artistic styles.
How Fooocus Image Generation Skill Works
- The skill initializes by performing an environment check to verify Python 3.10+, CUDA availability, and local Fooocus installation.
- It triggers the Fooocus service using context-specific scripts such as run_anime.bat for 2D styles or run_realistic.bat for photos.
- The AI agent interfaces with the Fooocus Gradio API or Web UI to input prompts, negative prompts, and style selections.
- For image-to-image tasks, the skill facilitates the upload of reference images and configuration of weights like PyraCanny or CPDS.
- During generation, the skill monitors real-time progress via WebSockets and provides updates until the final image is rendered.
- Generated assets and their associated metadata are saved to a structured local directory for easy retrieval.
Fooocus Image Generation Skill Setup
Prerequisites
Ensure you have an NVIDIA GPU with at least 4GB VRAM and 10GB of free disk space.
Installation Steps
# 1. Verify your system environment
python "${SKILL_ROOT}/scripts/check_env.py"
# 2. Install Fooocus and download base models
python "${SKILL_ROOT}/scripts/install_fooocus.py" --path ~/Fooocus --download-models
# 3. Start the service (Manual first-time recommended)
cd ~/Fooocus
python entry_with_update.py
Wait for the local URL (http://127.0.0.1:7865) to become active before sending requests through Openclaw Skills.
Fooocus Image Generation Skill Data Schema & Taxonomy
The skill manages data across several key directories within the Fooocus root:
| Component | Path | Description |
|---|---|---|
| Outputs | outputs/YYYY-MM-DD/ |
Daily folders containing generated PNG/WebP images and logs. |
| Checkpoints | models/checkpoints/ |
Storage for base SDXL models (.safetensors). |
| LoRAs | models/loras/ |
Directory for fine-tuned style or character models. |
| Inpaint | models/inpaint/ |
Specialized models for local image editing and repair. |
| Metadata | Embedded in PNG | Contains prompt, seed, styles, and performance settings for replication. |
Fooocus Image Generation Skill Advanced Features
- Multi-Style Blending: Combine over 200 built-in styles (e.g., SAI Anime + MRE Anime) to achieve unique artistic results.
- Advanced Control Engines: Utilize PyraCanny for edge control and CPDS for structural preservation during style migration.
- Metadata Extraction: Use the Describe feature to reverse-engineer prompts from existing images.
- Developer Debug Mode: Access granular parameters like Sampler type, Scheduler (Karras), and Guidance Scale for power-user optimization within Openclaw Skills.
- Real-time Upscaling: High-speed 1.5x and 2x upscaling using optimized SDXL refiner workflows.
SKILL.md
Loading
Related Openclaw Skills
Lead
A high-discipline decision skill designed to qualify sales prospects, identify information gaps, and recommend the next best action for commercial leads.
CronForge — Visual Cron Job Builder
A comprehensive library of human-readable cron job templates designed to eliminate the complexity of scheduling automation within the OpenClaw ecosystem.
Create Release (Gitee)
An automated release management skill for Gitee that generates semantic changelogs from pull request history and publishes structured releases.
Defense
An AI-powered legal defense skill designed to protect your rights and guide you through criminal, civil, and regulatory proceedings.
A/B Test Setup
A comprehensive toolkit for designing, implementing, and analyzing statistically rigorous A/B tests to drive data-informed product improvements.
Greece
An expert AI planning agent for navigating Greek island-hopping, mainland logistics, and seasonal travel safety.
Featured*
