🦉 Claw-Parade - Personal Brand Agent

AI-powered personal brand management system that transforms raw inputs into polished, multi-channel social media content with branded images and videos.

Transform articles, notes, ideas, and meeting recordings into ready-to-publish content across Instagram, LinkedIn, Twitter, and more — all while maintaining perfect brand consistency.

---

🎯 Features

Feature	💡 Benefit
Brand Consistency	Maintains your unique voice, tone, and visual identity across all content
Multi-Channel Output	Generates platform-specific content for Instagram, LinkedIn, Twitter, and more
Media Generation	Automated image and video processing with brand-aligned visuals
Smart Scheduling	Buffer-based scheduling with optimal posting times
Modular Skills	Extensible skill system for custom workflows
Docker Ready	Containerized deployment with GPU acceleration

---

🚀 Quick Start

Prerequisites

• Git (for submodule management)
• Docker (for containerized deployment)
• Python with uv (for AI enhancement tools)

Initialize Project

# Clone the repository
git clone <repo-url>
cd claw-parade

# Initialize skills as submodules
git submodule update --init --recursive

# Build Docker images
make build-base
docker build -t clawparade .

Run the Agent

# Initialize brand from your content
docker run -v $(pwd)/input:/input clawparade claw init --input /input/

# Process content for a channel
docker run -v $(pwd)/input:/input -v $(pwd)/output:/output clawparade \
  claw process --input /input/article.md --channel instagram

# Schedule content with a 5-day buffer
docker run -v $(pwd)/output:/output clawparade \
  claw schedule --buffer-days 5

---

📋 Workflow

Phase 1: Brand Initialization (One-Time)

Establish your brand identity by analyzing your content, resume, and past posts. This generates BRAND.md with:

• Voice and tone guidelines
• Visual identity specifications
• Core values and messaging

claw init --input ./raw-inputs/

Phase 2: Content Processing (Regular Use)

# Process a single file
claw process --input ./input/article.md --channel instagram

# Process video content
claw process --input ./input/meeting.mp4 --channel linkedin --video

# Bulk process from folder
claw process --folder ./input/ --channel twitter

# Schedule with buffer
claw schedule --buffer-days 5

---

📁 Directory Structure

claw-parade/
├── README.md              # This file
├── SOUL.md               # Agent identity and behavior specification
├── BRAND.md              # Generated brand spec (created by init)
├── WORKFLOW.md           # Detailed workflow documentation
├── Dockerfile           # Container build configuration
├── docker-compose.yml   # Container orchestration
├── skills/              # Modular skill definitions
│   ├── brand-awareness/  # Brand identity maintenance
│   ├── video-processing/  # Video enhancement and captioning
│   ├── image-processing/  # Image resize and filtering
│   ├── buffer/           # Schedule and publish posts
│   ├── + 10 standalone skills # AI enhancement tools
│   └── SKILL.md files    # Individual skill documentation
├── input/               # Raw input files (articles, notes, ideas)
└── output/              # Processed content organized by channel
    ├── instagram/
    ├── linkedin/
    └── twitter/

---

🔧 Docker Integration

Docker Compose

# Run as service
docker compose up openclaw-gateway

# Run CLI commands
docker compose run openclaw-cli claw init --input ./input/

Environment Variables

# Core configuration
CLAW_BRAND_FILE=./BRAND.md         # Path to brand specification
CLAW_INPUT_DIR=./input/            # Raw input location
CLAW_OUTPUT_DIR=./output/          # Processed output location

# Scheduling defaults
CLAW_BUFFER_DAYS=5                 # Default buffer days
CLAW_DEFAULT_CHANNEL=instagram     # Default target channel

# External services
CLAW_GDRIVE_ENABLED=true           # Enable Google Drive sync
BUFFER_API_KEY=your-buffer-token   # Required for scheduling

---

🎨 AI Enhancement Tools

OpenClaw includes 12 standalone AI enhancement tools that can be used independently:

Tool	Input	What it does	Min VRAM
clawaes	images	Score and pick the best photos	~1 GB
clawdepth	images	Synthetic bokeh / portrait mode	~1.5 GB
clawbg	images	Remove background, replace with colour/image	~0.5 GB
clawvlm	images	Auto-describe and suggest captions	~4 GB
clawrife	videos	Frame interpolation (60fps / slow motion)	~2 GB
clawmatte	videos	Remove video background, composite backdrop	~3 GB
clawsep	video/audio	Separate vocals from music	~2 GB
clawbeat	prompt / video	Generate brand background music	~3 GB
clawanimate	images	Image → animated video clip	~8 GB
clawvace	video	Edit / inpaint video regions via prompt	~8 GB
clawportrait	portrait + driver video	Animate a face from a driving video	~4 GB

Usage: Drop files into ./input, get results in ./output. Each tool follows the same conventions (uv sync, --input, --output, --device cpu fallback).

---

📚 Documentation

Document	Description
🏷️ SOUL.md	Agent identity, persona, and behavior specs
📖 WORKFLOW.md	Complete processing workflow and best practices
🔧 Skills	Individual skill documentation and configuration
🎨 AI Enhancements	Detailed AI tool documentation (notes)

---

🚀 Getting Help

CLI Commands

# Get help for any command
claw --help
claw init --help
claw process --help
claw schedule --help

Common Issues

Issue	Resolution
No BRAND.md	Auto-trigger Init phase
Unclear input	Request clarification via chat
Media generation fail	Fall back to templates
Scheduling fail	Queue to manual review
Channel not supported	Suggest alternative channels

---

📄 License

MIT

---