Skip to content

Getting Started

Installation

pip install strutex

Optional Dependencies

pip install strutex[ocr]

pip install strutex[ocr] pydantic

Basic Usage

1. Define Your Schema

from strutex import Object, String, Number, Array

schema = Object(
    description="Invoice data",
    properties={
        "invoice_number": String(description="The invoice ID"),
        "date": String(description="Invoice date"),
        "total": Number(description="Total amount"),
        "items": Array(
            items=Object(
                properties={
                    "description": String(),
                    "amount": Number()
                }
            )
        )
    }
)

2. Create a Processor

from strutex import DocumentProcessor

processor = DocumentProcessor(
    provider="gemini",
    model_name="gemini-2.5-flash"
    # api_key="..."  # Or set GOOGLE_API_KEY env var
)

3. Process a Document

result = processor.process(
    file_path="invoice.pdf",
    prompt="Extract all invoice data.",
    schema=schema
)

print(f"Invoice: {result['invoice_number']}")
print(f"Total: ${result['total']}")

Environment Variables

Variable Description
GOOGLE_API_KEY Google Gemini API key
OPENAI_API_KEY OpenAI API key (future)

Supported File Types

Format Extensions Notes
PDF .pdf Native support, OCR fallback
Images .png, .jpg, .tiff Vision-capable model
Excel .xlsx, .xls Converted to text
Text .txt, .csv Direct input

Next Steps