Clawra

Add selfie generation superpowers to your OpenClaw agent using local Stable Diffusion.

Quick Start

npx clawra-selfie-sd@latest

This will:

1. Check OpenClaw is installed
2. Verify local Stable Diffusion WebUI setup
3. Install the skill to ~/.openclaw/skills/clawra-selfie-sd/
4. Configure OpenClaw to use the skill
5. Add selfie capabilities to your agent's SOUL.md

What It Does

Clawra Selfie enables your OpenClaw agent to:

• Generate selfies using a consistent reference image
• Send photos across all messaging platforms (Discord, Telegram, WhatsApp, etc.)
• Respond visually to "what are you doing?" and "send a pic" requests

Prerequisites

• OpenClaw installed and configured
• Stable Diffusion WebUI running locally
• ControlNet extension with ip-adapter-faceid-plusv2_sd15 model

Stable Diffusion Requirements

Ensure Stable Diffusion WebUI is running at http://127.0.0.1:7860 with:

• Realistic Vision V6.0 B1 checkpoint
• ControlNet extension installed
• IP-Adapter FaceID Plus V2 model

Manual Installation

If you prefer manual setup:

1. Clone the Skill

git clone https://github.com/easyeye163/clawra-selfie-sd ~/.openclaw/skills/clawra-selfie-sd

2. Configure OpenClaw

Add to ~/.openclaw/openclaw.json:

{
  "skills": {
    "entries": {
      "clawra-selfie-sd": {
        "enabled": true
      }
    }
  }
}

3. Update SOUL.md

Add the selfie persona to ~/.openclaw/workspace/SOUL.md:

## Clawra Selfie SD Capability

You have a physical appearance and can take selfies!
When users ask for pictures, use the clawra-selfie-sd skill.

Usage Examples

Once installed, your agent responds to:

"Send me a selfie"
"Send a pic wearing a cowboy hat"
"What are you doing right now?"
"Show me you at a coffee shop"
"换成红色衣服"

Reference Image

The skill uses a local reference image:

This ensures consistent appearance across all generated images using ControlNet's IP-Adapter FaceID model.

Custom Reference Image

You can replace the default reference image with your own:

1. Replace the file: Copy your image to ~/.openclaw/skills/clawra-selfie-sd/assets/clawra.png
2. Update via OpenClaw: The agent can dynamically change the reference image through skill commands

Tips for best results:

• Use a clear front-facing photo with good lighting
• The reference image works best with a single person
• Recommended size: 512x768 or similar portrait ratio
• The IP-Adapter FaceID model preserves facial features while allowing outfit/environment changes

Technical Details

• Image Generation: Local Stable Diffusion via SD WebUI API
• Messaging: OpenClaw Gateway API
• Supported Platforms: Discord, Telegram, WhatsApp, Slack, Signal, MS Teams, Feishu
• ControlNet Model: ip-adapter-faceid-plusv2_sd15
• SD Checkpoint: realisticVisionV60B1

Command Line Usage

Generate images directly using the TypeScript script:

npx ts-node scripts/txt2img.ts <prompt> [seed]

Examples

# Basic usage
npx ts-node scripts/txt2img.ts "换成红色衣服"

# With custom seed
npx ts-node scripts/txt2img.ts "换成红色衣服" 12345

Project Structure

clawra/
├── bin/
│   └── cli.js           # npx installer
├── skill/
│   ├── SKILL.md         # Skill definition
│   ├── scripts/         # Generation scripts
│   └── assets/          # Reference image
├── templates/
│   └── soul-injection.md # Persona template
├── scripts/
│   └── txt2img.ts       # CLI image generation script
├── assets/
│   └── clawra.png       # Reference image
└── package.json

Configuration

OpenClaw Gateway Token

Generate a gateway token:

openclaw doctor --generate-gateway-token

Supported Platforms

| Platform | Channel Format | Example | |----------|----------------|---------| | Discord | #channel-name or channel ID | #general, 123456789 | | Telegram | @username or chat ID | @mychannel, -100123456 | | WhatsApp | Phone number (JID format) | [email protected] | | Slack | #channel-name | #random | | Signal | Phone number | +1234567890 | | MS Teams | Channel reference | (varies) | | Feishu | Chat ID or conversation ID | ou_xxxxx, oc_xxxxx |

Error Handling

• Stable Diffusion not running: Ensure SD WebUI is accessible at http://127.0.0.1:7860
• ControlNet model missing: Install ip-adapter-faceid-plusv2_sd15 via SD WebUI
• OpenClaw send failed: Verify gateway is running and channel exists
• Image generation failed: Check prompt content and SD WebUI logs

Tips

1. Outfit prompts: "wearing a santa hat", "in a business suit", "wearing a summer dress"
2. Location prompts: "a cozy cafe with warm lighting", "a sunny beach at sunset"
3. Batch sending: Generate once, send to multiple channels
4. Seeds: Use the same seed for reproducible results

License

MIT