Turn any idea into a scroll-stopping AI video, powered by Google's VEO-3, no design skills required.

VeoCrafter is an automated video generation pipeline that transforms simple text ideas into engaging short-form videos using Google's VEO-3 AI model. The system handles everything from concept brainstorming to final video rendering, making viral content creation accessible to everyone.

• Automated Idea-to-Video Pipeline: Input a topic and receive multiple polished video concepts with finished clips
• Intelligent Concept Generation: GPT-4.1 agent creates viral-ready hooks, captions, and video settings
• Smart Prompt Optimization: Dedicated agent refines concepts into VEO-3-optimized prompts
• Seamless VEO-3 Integration: Direct access through fal.ai endpoint without expensive Google contracts
• Automated Workflow: Batch processing with comprehensive Excel logging of all generated content

• Social Media Content: Generate Reels, Shorts, and TikToks at scale
• Marketing Campaigns: Test creative concepts before investing in production
• Rapid Prototyping: Develop storyboards and visual concepts quickly
• Educational Content: Create engaging explainer videos for complex topics
• Creative Exploration: Experiment with unique video concepts and styles

Topic Input → Concept Agent → Prompt Agent → VEO-3 Rendering → Excel Logging
     ↓              ↓             ↓              ↓              ↓
   "Alien food   Viral hooks   Optimized     MP4 video     Metadata
    critic"      & captions    VEO prompts    generation     tracking

1. Topic Input: Provide a simple text description of your video concept
2. Concept Generation: GPT-4.1 creates multiple engaging video ideas with hooks and captions
3. Prompt Optimization: Second agent converts concepts into precise VEO-3 prompts
4. Video Rendering: fal.ai processes requests through Google's VEO-3 model
5. Result Logging: All videos, prompts, and metadata automatically saved to Excel

veo-crafter/
├── main.py              # Main orchestration script
├── prompts.py           # System prompts and templates
├── utils.py             # API utilities and Excel export
├── requirements.txt     # Python dependencies
├── .env                 # Environment variables (not tracked)
└── videos.xlsx          # Generated video log (auto-created)

# Clone the repository
git clone https://github.com/prajvalrasik/veo-crafter.git
cd ad_generation_agent

# Create virtual environment (recommended)
python -m venv venv

# Activate virtual environment
# Windows:
.\venv\Scripts\Activate.ps1
# macOS/Linux:
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

Create a .env file in the project root:

FAL_KEY=your_fal_api_key_here
OPENROUTER_API_KEY=your_openrouter_key_here

Edit main.py to set your desired topic and video count:

topic = "Alien food critic reviews Earth cuisine"
count = 3  # Number of videos to generate

Run the generator:

python main.py

Generated videos and metadata will be saved in videos.xlsx. Each row contains:

• Video URL
• Original prompt
• Caption
• Timestamp
• Generation parameters

• Cost: Approximately $0.75 per second of video via fal.ai
• Duration Limit: 8 seconds maximum per video
• Rate Limits: Subject to fal.ai and VEO-3 API limitations

• fal.ai Documentation
• Google VEO-3 Information
• OpenRouter API

Made with ❤️