MovieStudio

AI generated movie video

README.md

AI Movie Generator with CrewAI

An open-source automated movie production system that transforms story prompts into complete short films using CrewAI agents. Based on the revolutionary concepts from "Tomorrow's Film Studio Today", this system orchestrates multiple AI agents to handle every aspect of film production - from scriptwriting to visual development, character design, cinematography, and video generation preparation.

🎬 Features

Automated Screenplay Writing: Transforms story concepts into professionally formatted scripts
Visual Development Pipeline: Generates concept art, mood boards, and storyboards using DALL-E 3
Virtual Actor Creation: Designs detailed character sheets with multiple angles and expressions
Cinematography Planning: Creates comprehensive shot lists with camera angles and movements
Sound Design Architecture: Develops detailed audio landscapes including dialogue, foley, and music
Video Generation Ready: Produces prompts optimized for Runway Gen-4, Google Veo, and other AI video tools
Modular Architecture: Easy to extend with new agents and capabilities
Multiple Style Presets: Cinematic, documentary, noir, and customizable styles

🏗️ System Architecture

The system employs six specialized CrewAI agents working in sequence:

1. Screenplay Writer Agent

Transforms story prompts into professional screenplay format
Handles three-act structure, character development, and dialogue
Outputs industry-standard formatted scripts

2. Visual Development Artist Agent

Creates concept art for key scenes
Establishes visual style and mood
Generates color palettes and lighting references

3. Character Designer Agent

Designs virtual actors with full character sheets
Creates multiple angle views and expression guides
Develops costume and prop details

4. Virtual Cinematographer Agent

Plans shot compositions and camera movements
Creates detailed shot lists with technical specifications
Designs visual flow and pacing

5. Sound Designer Agent

Develops comprehensive sound design documents
Plans dialogue, foley, ambience, and music
Creates cue sheets and audio specifications

6. AI Video Director Agent

Synthesizes all elements into video generation prompts
Optimizes for specific AI video platforms
Maintains consistency across all generated content

📋 Prerequisites

Python 3.8 or higher
OpenAI API key (for GPT-4 and DALL-E 3)
FFmpeg (for video processing)
8GB+ RAM (16GB recommended)
10GB+ free disk space
Internet connection for API calls

🚀 Installation

Clone the repository:

git clone https://github.com/yourusername/ai-movie-generator.git
cd ai-movie-generator

Create a virtual environment (recommended):

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

Install dependencies:

pip install crewai crewai-tools langchain-openai openai requests pillow

Install FFmpeg:

macOS: brew install ffmpeg
Ubuntu/Debian: sudo apt-get install ffmpeg
Windows: Download from ffmpeg.org

Set up environment variables:

export OPENAI_API_KEY="your-openai-api-key-here"

Or create a .env file:

OPENAI_API_KEY=your-openai-api-key-here

💡 Usage

Basic Usage

Generate a complete movie production from a simple prompt:

python main.py "A lonely robot discovers an abandoned flower in a post-apocalyptic city"

Advanced Usage

Customize style, duration, and output:

python main.py "Your story prompt" --style noir --duration 300 --output my_movie.mp4

Command Line Options

prompt (required): Your story prompt
--style: Visual style preset (default: cinematic)
- cinematic: Professional film look with dramatic lighting
- documentary: Natural, observational style
- noir: High contrast, moody atmosphere
--duration: Target duration in seconds (default: 180)
--output: Output filename (default: my_movie.mp4)

Example Commands

# Generate a 5-minute sci-fi film
python main.py "An AI gains consciousness during a solar storm" --duration 300

# Create a noir-style mystery
python main.py "A detective investigates dreams that predict crimes" --style noir

# Short experimental film
python main.py "Colors have sounds in a synesthetic world" --duration 120 --output synesthesia.mp4

📁 Project Structure

ai-movie-generator/
├── movie_generator.py      # Core CrewAI agents and tasks
├── config.py              # Configuration and settings
├── video_assembler.py     # Video assembly and processing tools
├── main.py                # Command-line interface
├── movie_template.md      # Screenplay formatting template
├── requirements.txt       # Python dependencies
├── .env.example          # Environment variable template
├── output/               # Generated content directory
│   ├── scripts/         # Generated screenplays
│   ├── concept_art/     # Visual development artwork
│   ├── storyboards/     # Shot-by-shot storyboards
│   ├── characters/      # Character design sheets
│   ├── production/      # Shot lists and production documents
│   │   └── shot_lists/  # Detailed shot breakdowns
│   ├── sound/          # Sound design documents
│   ├── video_prompts/  # AI video generation prompts
│   └── final/          # Final assembled videos
└── README.md           # This file

📦 Generated Outputs

1. Screenplay (`output/scripts/`)

Professional screenplay format
Scene headings (INT./EXT.)
Action lines and dialogue
Character introductions
Format: .fountain and .pdf

2. Concept Art (`output/concept_art/`)

High-resolution concept images (1792x1024)
Key scene visualizations
Mood and atmosphere references
Color palette explorations

3. Storyboards (`output/storyboards/`)

Shot-by-shot visual planning
Black and white sketch style
Camera angle indicators
Movement and transition notes

4. Character Designs (`output/characters/`)

Multi-angle character sheets
Expression and emotion guides
Costume and prop details
Size comparison charts

5. Production Documents (`output/production/`)

shot_list.json: Complete shot breakdown
camera_plan.md: Cinematography notes
schedule.json: Production timeline
tech_specs.md: Technical requirements

6. Sound Design (`output/sound/`)

sound_design.json: Complete audio plan
dialogue_notes.md: Voice direction
foley_list.csv: Required sound effects
music_brief.md: Score direction

7. Video Prompts (`output/video_prompts/`)

prompt_[shot_number].json: Individual shot prompts
style_guide.json: Consistent visual parameters
transition_guide.md: Shot connections
vfx_requirements.json: Special effects needs

⚙️ Sample OUTPUT

Creating a movie about the Dust Bowl: "So Shall Ye Reap"

The Script Emerges

The opening scene went through seven AI iterations before finding its voice:

FADE IN: EXT. HADLEY FARM - DAWN - 2040 What was once 300 acres of cotton now resembles Mars. Red dust swirls in heat shimmer despite the early hour. A digital thermometer on a fence post reads 97°F at 5:47 AM.

Farm

SARAH HADLEY (73), leather-faced but unbowed, stands in what used to be her garden. She holds a handful of dust, lets it slip through her fingers.

Sarah

SARAH (to the dust) Sixty years I fed you. Compost. Manure. Love. This is how you repay me? A screen door SLAMS. TOM HADLEY (45), farmer's build gone soft from worry, joins her. His AgriTech uniform is three years out of date.

EMMA Mama, you need to come inside. Air quality's red again.

Emma

SARAH Air quality's been red since I can last remember. Tom's smartwatch CHIRPS. He glances at it, his face tightening.

VEO Movie clip "So Shall Ye Reap"

Watch the video

Modify Style Presets

Edit config.py to add custom styles:

STYLE_PRESETS = {
    "custom_style": {
        "lighting": "your lighting description",
        "color": "your color palette",
        "camera": "your camera style",
        "mood": "your mood description"
    }
}

Adjust Video Settings

VIDEO_SETTINGS = {
    "resolution": "1920x1080",  # or "3840x2160" for 4K
    "fps": 24,                  # or 30, 60
    "aspect_ratio": "16:9",     # or "21:9", "4:3"
    "codec": "h264",
    "bitrate": "10M"
}

Add Shot Types

SHOT_TYPES = {
    "custom_shot": {
        "abbr": "CS",
        "description": "Your shot description"
    }
}

🔧 Extending the System

Adding New Agents

Create a new agent in movie_generator.py:

new_agent = Agent(
    role='Your Agent Role',
    goal='What this agent accomplishes',
    backstory="Agent's background and expertise",
    verbose=True,
    llm=llm,
    tools=[your_custom_tools],
    allow_delegation=False
)

Creating Custom Tools

Add new tools with the @tool decorator:

@tool
def your_custom_tool(parameter: str) -> str:
    """
    Description of what your tool does.
    Used by agents for specific tasks.
    """
    # Your implementation
    result = process_something(parameter)
    return result

Integrating Video Generation APIs

When Runway Gen-4 or similar APIs become available:

@tool
def generate_video_with_runway(prompt: str, duration: int) -> str:
    """Generate video using Runway Gen-4 API"""
    # API implementation
    return video_path

🎥 Workflow Overview

Story Input → Script Writer Agent creates screenplay
Script Analysis → Visual Developer creates concept art
Character Extraction → Character Designer creates actors
Scene Breakdown → Cinematographer plans shots
Audio Planning → Sound Designer creates soundscape
Synthesis → Video Director creates generation prompts
Assembly → Video tools compile final movie

⚡ Performance Optimization

Batch Processing: Generate multiple images simultaneously
Caching: Reuse generated content across similar projects
Resolution Options: Lower resolution for drafts, high for finals
Selective Generation: Only regenerate changed elements

🐛 Troubleshooting

Common Issues

API Rate Limits
- Solution: Add delays between API calls
- Use exponential backoff for retries
Memory Errors
- Solution: Reduce image resolution
- Process in smaller batches
FFmpeg Not Found
- Solution: Ensure FFmpeg is in system PATH
- Verify installation: ffmpeg -version
API Key Errors
- Solution: Check API key is correctly set
- Verify quota hasn't been exceeded

Debug Mode

Run with verbose output:

python main.py "Your prompt" --debug

🤝 Contributing

We welcome contributions! Here's how to help:

Areas for Improvement

Video Generation Integration: Connect with Runway, Pika, etc.
Audio Generation: Integrate ElevenLabs, Suno
Character Animation: Add movement and expression systems
Real-time Preview: Build preview system during generation
Multi-language Support: Internationalization
GUI Interface: Create user-friendly interface
Cloud Deployment: Dockerize and deploy to cloud

Contribution Process

Fork the repository
Create feature branch (git checkout -b feature/amazing-feature)
Commit changes (git commit -m 'Add amazing feature')
Push to branch (git push origin feature/amazing-feature)
Open Pull Request

Code Style

Follow PEP 8
Add docstrings to all functions
Include type hints
Write unit tests for new features

📈 Roadmap

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

Built with CrewAI
Inspired by "Tomorrow's Film Studio Today" by H. Peter Alesso
Uses OpenAI's GPT-4 and DALL-E 3
Community contributors and testers

📞 Contact & Support

GitHub Issues: For bug reports and feature requests
Discussions: For general questions and ideas
Email: [email protected]
Discord: Join our community

💰 Costs

Estimated costs per movie (3-5 minutes):

Script Generation: ~$0.50-$1.00
Concept Art (10 images): ~$2.00-$3.00
Storyboards (20-30 panels): ~$4.00-$6.00
Character Designs (3-5 characters): ~$1.00-$1.50
Total: ~$7.50-$11.50 per short film

Note: Actual video generation costs (Runway, etc.) not included as APIs are not yet publicly available.

🚦 Quick Start Guide

Install: pip install -r requirements.txt
Configure: Set your OpenAI API key
Run: python main.py "A robot learns to paint"
Wait: 10-20 minutes for full generation
Explore: Check output/ directory for results

⚠️ Important Note: This system currently generates all production documents, scripts, artwork, and prompts needed for AI video generation. Actual video synthesis requires access to AI video generation services (Runway Gen-4, Google Veo, Pika, etc.) which may have separate costs, availability requirements, and API access. The system is designed to seamlessly integrate with these services as they become publicly available.

CONTACT

Homepage: Video Software Lab

Email: [email protected]

GitHub: [https://github.com/alessoh/MovieStudio]

For any questions, feedback, or bug reports, please open an issue in the GitHub repository or contact us via email

Acknowl Last Updated: December 2024
Version: 1.0.0

Name		Name	Last commit message	Last commit date
Latest commit History 9 Commits
__pycache__		__pycache__
characters		characters
concept_art		concept_art
production		production
sound		sound
video_prompts		video_prompts
.gitignore		.gitignore
Emma1.png		Emma1.png
LICENSE		LICENSE
README.md		README.md
config.py		config.py
farm1.png		farm1.png
main.py		main.py
movie_generator.py		movie_generator.py
movie_template.md		movie_template.md
requirements.txt		requirements.txt
sarah2.png		sarah2.png
video_assembler.py		video_assembler.py
video_generation_guide.md		video_generation_guide.md

License

alessoh/MovieStudio

Folders and files

Latest commit

History

Repository files navigation