It's a forge - where you choose your materials (LLM, TTS, image generators), heat them in your fire (configuration), and hammer them into your perfect video creation engine.

vs Traditional Tools

Traditional Tools        ReelForge
━━━━━━━━━━━━━━━        ━━━━━━━━━━━━
Fixed workflow      →   Modular & Customizable
One-size-fits-all   →   Forge Your Own
Closed system       →   Open & Extensible

⚒️ Core Philosophy

Every creator forges differently.

🎯 Modular by Design - Swap TTS, image gen, frame templates
🔧 Plug & Play - Use built-in or bring your own
🎨 DIY Workflows - Customize every step of generation
🌐 Open Source - Community-driven and transparent

🚀 Quick Start

Installation

# Clone the repository
git clone https://github.com/YOUR_USERNAME/ReelForge.git
cd ReelForge

# Install dependencies with uv
uv sync

# Copy config example
cp config.example.yaml config.yaml

# Edit config.yaml and fill in your API keys

Your First Forge

# Run the web interface
uv run python web.py

# Or use CLI
uv run reelforge --help

Programmatic Usage

from reelforge.service import reelforge

# Initialize
await reelforge.initialize()

# Generate a video
result = await reelforge.book_video_workflow.generate(
    book_name="Atomic Habits",
    n_storyboard=5
)

print(f"Forged: {result.video_path}")

🔨 What Makes ReelForge Different?

1. Truly Modular Architecture

Swap components like changing tools in your forge:

# config.yaml
llm:
  api_key: your_key
  model: qwen-max

tts:
  default: edge  # or azure, elevenlabs

image:
  default: comfykit
  comfykit:
    comfyui_url: http://localhost:8188

2. Capability-Based System

Every capability follows a simple naming convention:

{type}_{id}

Examples:
  llm_qwen         → LLM capability, ID: qwen
  tts_edge         → TTS capability, ID: edge
  image_comfykit   → Image capability, ID: comfykit

3. Storyboard-Based Generation

Unlike simple template filling, ReelForge uses AI to:

Understand your topic deeply
Generate narrative storyboards
Create scene-by-scene visuals
Compose professional videos

🏗️ Architecture

ReelForge Core
├── Capabilities (Pluggable)
│   ├── LLM (OpenAI, Qwen, Ollama, ...)
│   ├── TTS (Edge, Azure, ElevenLabs, ...)
│   ├── Image (ComfyUI, SD, DALL-E, ...)
│   └── Custom (Your own!)
├── Services (Composable)
│   ├── Narration Generator
│   ├── Image Prompt Generator
│   ├── Storyboard Processor
│   ├── Frame Composer
│   └── Video Compositor
└── Workflows (Customizable)
    └── Define your own pipeline

🎨 Built-in Capabilities

LLM (Large Language Models)

Provider	ID	API Key Required
通义千问	`qwen`	✅ DASHSCOPE_API_KEY
OpenAI GPT	`openai`	✅ OPENAI_API_KEY
DeepSeek	`deepseek`	✅ DEEPSEEK_API_KEY
Ollama	`ollama`	❌ No (runs locally)

TTS (Text-to-Speech)

Provider	ID	Features
Edge TTS	`edge`	Free, Multiple voices
Azure TTS	`azure`	Premium quality

Image Generation

Provider	ID	Features
ComfyKit	`comfykit`	Custom workflows, Local
(More coming)	-	-

📚 Examples

Check out the examples/ directory:

generate_video_simple.py - Basic video generation
generate_video_with_bgm.py - Add background music
generate_video_with_image_style.py - Custom image styles
generate_video_custom.py - Full customization

🤝 Contributing

ReelForge is community-driven.

🐛 Report bugs
💡 Request features
🔧 Build components
🎨 Share your forge configurations

🗺️ Roadmap

Core modular architecture
Built-in LLM/TTS/Image capabilities
Storyboard-based generation
Web interface (Streamlit)
Visual forge builder
Community forge marketplace
Plugin ecosystem
Cloud hosting

📄 License

MIT License - see LICENSE

🙏 Acknowledgments

Built with ❤️ by the creator community.

Inspired by:

Forge Your Way
Every creator forges differently. What will you forge?

Star us if you like what we're building! ⭐

Languages

Python 73%

HTML 16.8%

TypeScript 8.9%

Shell 0.6%

CSS 0.3%

Other 0.4%