vision-ocr

vision-ocr

Convert images to markdown using Llama 3.2 Vision model powered by Groq AI.

Features

• 📸 Convert images to clean, formatted markdown
• 🌐 Support for both local and remote images
• 🎯 Accurate text extraction and formatting
• 📝 Preserves document structure and formatting
• ⚡ Fast processing using Groq AI

Installation

npm install vision-ocr

Setup

• Get your Groq API key from Groq Console
• Set up your API key either:
  • As an environment variable: GROQ_API_KEY=your_api_key
  • Or pass it directly in the code

Usage

import { ocr } from 'vision-ocr';

// Using a remote image
const markdown = await ocr({
  filePath: 'https://example.com/image.jpg',
  apiKey: 'your_groq_api_key' // Optional if env variable is set
});

// Using a local image
const markdown = await ocr({
  filePath: './path/to/local/image.jpg'
});

console.log(markdown);

API Reference

ocr(options)

Main function to convert images to markdown.

Options

Parameter	Type	Required	Default	Description
filePath	string	Yes	-	Path to local image or URL of remote image
apiKey	string	No	process.env.GROQ_API_KEY	Your Groq API key
model	string	No	"llama-3.2-11b-vision-preview"	Model to use for OCR

Supported Models

• llama-3.2-11b-vision-preview
• llama-3.2-90b-vision-preview

Error Handling

The package throws errors in these cases:

• Missing file path
• Invalid image file
• API errors
• No content extracted from image

try {
  const markdown = await ocr({
    filePath: 'path/to/image'
  });
} catch (error) {
  console.error('OCR failed:', error.message);
}

Limitations

• Images must be in a supported format (JPEG, PNG)
• Maximum image size depends on Groq API limits
• API key required for operation

License

MIT © Saurabh Udupi (@0xSaurabhx)

Contributing

Contributions, issues, and feature requests are welcome! Feel free to check the issues page.

Credit

This project was inspired by nutlope. Go check it out!