Text Extraction with OpenAI Vision API in a Node.js Application

A highly optimized Node.js application demonstrating cost-effective text extraction from images using OpenAI's Vision API. This project showcases how to achieve significant cost reductions through model selection and token optimization while maintaining high accuracy.

🎯 Purpose

This project demonstrates:

• How to optimize token usage in OpenAI's Vision API
• Best practices for performance monitoring
• Efficient image processing strategies

💡 Key Findings

Token Optimization

• Optimized token usage: ~110 tokens per image
• Achieved through model selection (GPT-4 Turbo) (maybe use gpt-4o-mini model, because more capable for text extraction)
• Optimized system prompts and messages
• Implemented "low detail" mode without accuracy loss

Cost Efficiency

• Cost per image: $0.0012 (0.04 TL)
• 96% cost reduction achieved
• Highly scalable for large datasets

🚀 Features

• Efficient text extraction using GPT-4 Turbo
• Detailed token usage analytics
• Timestamped results with comprehensive stats

📊 Performance Metrics

• Token usage: ~110 per image
• Cost per image: $0.0012 (0.04 TL)
• Processing time: 1-2 seconds
• Accuracy rate: Very High

💰 Cost Analysis (Optimized)

Scale comparison at current API prices:

• 1,000 images: ~$1.20 (42 TL)
• 10,000 images: ~$12.00 (420 TL)
• 100,000 images: ~$120.00 (4,200 TL)

🛠️ Setup & Usage

1. Clone the repository:

git clone https://github.com/yourusername/openai-vision-optimizer.git
cd openai-vision-optimizer

2. Install dependencies:

npm install

3. Create a .env file in the root directory:

OPENAI_API_KEY=your_api_key_here

4. Place your PNG files in the /image folder

5. Run the application:

node captcha_solver.js

Results are automatically saved in /results with timestamps (format: results-YYYYMMDD-HHMMSS.txt)

📝 Output Format

filename.png: extracted_text
  Response Time: X.XX seconds
  Token Usage: XXX (XXX input + XX output)
  Cost: $X.XXXX (X.XX TL)

⚙️ Technical Optimizations

1. API Optimization

   • Minimal system prompts
   • Optimized token usage
   • Rate limit handling

2. Cost Management

   • Token usage monitoring
   • Detailed reporting
   • Multi-currency support (USD/TL)

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

⚠️ Disclaimer

This project is not affiliated with OpenAI. All API pricing and performance metrics are subject to change. Please refer to OpenAI's official documentation for current pricing.

📸 Example Captcha Images

Below are some example captcha images that can be processed using this application: