Quickstart

Extract structured data from your first document in 5 minutes.

Installation

pip install strutex

Your First Extraction

from strutex import DocumentProcessor
from strutex.schemas import INVOICE_US

# Create processor
processor = DocumentProcessor(provider="gemini")  # or "openai", "anthropic"

# Extract structured data
result = processor.process(
    file_path="invoice.pdf",
    prompt="Extract invoice details",
    schema=INVOICE_US
)

# Use the result
print(f"Vendor: {result['vendor_name']}")
print(f"Total: ${result['total']}")
print(f"Date: {result['date']}")

That's it! You've extracted structured JSON from a document.

---

What Just Happened?

1. DocumentProcessor — The main engine that handles extraction
2. INVOICE_US — A built-in schema defining what fields to extract
3. process() — Sends the document to an LLM and validates the result

---

Next Steps

Want to...	Go to...
Define your own schema	Your First Schema
Try different providers	Switching Providers
Add data validation	Adding Validation

---

Troubleshooting

"API key not found"

export GOOGLE_API_KEY="your-key"  # For Gemini
export OPENAI_API_KEY="your-key"  # For OpenAI

"File not found"

# Use absolute path
result = processor.process("/full/path/to/invoice.pdf", ...)