scanverion

Node and Python are the first supported Scanverion client languages. This package provides synchronous and asynchronous standard-library clients for the OCR API.

Download and install

Install the published package from PyPI:

python -m pip install scanverion==0.1.0

It requires Python 3.10 or newer.

Quickstart

from pathlib import Path
from scanverion import Scanverion

client = Scanverion(
    api_key="skv_...",
    base_url="http://localhost:5009",
)
result = client.documents.process(
    Path("invoice.pdf"),
    file_name="invoice.pdf",
    content_type="application/pdf",
    operations=["Detection", "Scanner"],
    scanner_fields=["invoiceNumber", "grossTotal"],
    idempotency_key="invoice-2026-0007",
)
print(result["requestId"], result["objects"])

The core methods accept bytes, bytearray, memoryview, pathlib.Path, or binary file objects. AsyncScanverion exposes the same resource groups and uses an injected async transport or asyncio.to_thread with the default transport.

from scanverion import AsyncScanverion

async with AsyncScanverion(api_key="skv_...") as client:
    summary = await client.usage.summary(from_date="2026-07-01", to_date="2026-07-31")

API keys are server-side credentials. Do not put them in browser code, mobile binaries, logs, or public fixtures. Processing POST retries happen only when an idempotency key is supplied; safe GET resources may retry transient failures.

Local validation

From the scanverion directory:

python -m unittest discover -s packages/sdk-python/tests -v
python -m compileall packages/sdk-python/src