ABBYY is introducing a brand new optical character recognition (OCR) API to allow builders to extract knowledge from unstructured paperwork.
“As a vanguard of OCR, ABBYY has lengthy had a vibrant group of cutting-edge builders creating transformational options with our superior doc AI,” mentioned Nick Hyatt, vice chairman of Engineering R&D at ABBYY. “ABBYY Doc AI API is a significant step ahead for creating automated doc workflows.”
The ABBYY Doc AI API—at present in technical preview—will enable builders to rework unstructured knowledge into structured JSON in only a few strains of code. It consists of SDKs for Python, C#, JavaScript, and Java.
Some examples of paperwork that knowledge will be transformed from embrace invoices, receipts, and tax types.
Throughout this technical preview, the OCR fashions are solely out there as pre-trained fashions, with no choices for customized coaching or fine-tuning but. The API might be free to make use of in the course of the preview, however there’s a processing quantity restrict of 1000 pages.
It at present helps OCR in English, German, French, Spanish, Dutch, Japanese, and each conventional and simplified Chinese language. For handwriting recognition, or ICR, it helps English, German, French, Spanish, and Japanese.
Builders can be a part of the technical preview right here.
