Zerox
StaleDescription
OCR and document extraction tool using vision models, efficiently converting PDFs and images into structured text.
Key Features
- Converts PDFs, DOCX, images to images then calls vision models to generate Markdown
- Supports OpenAI, Azure OpenAI, AWS Bedrock, Google Gemini, Vertex AI providers
- Dual SDK in Node.js and Python with async API and concurrent processing
- Built-in structured data extraction with custom Schema support
- Page orientation correction, edge trimming, and format preservation preprocessing
- maintainFormat option ensures consistent output for cross-page tables and complex layouts
Use Cases
💡 Convert PDF documents to searchable Markdown text in RAG pipelines
💡 Automatically extract structured information from invoices and contracts
💡 Batch process scanned files and generate editable text versions
💡 Convert chart-heavy and table-heavy documents into AI-understandable formats
💡 Provide document understanding capabilities for AI agents with multimodal input
Tags
Categories
Quick Start
1. Install Node.js SDK: `npm install zerox`
2. Usage example:
```ts
import { zerox } from 'zerox';
const result = await zerox({
filePath: 'path/to/document.pdf',
credentials: { apiKey: process.env.OPENAI_API_KEY },
});
```
3. Or install Python SDK: `pip install py-zerox`