Document OCR

The Actor takes list of document urls (PDFs or images) and language, runs them through OCR, and stores the results in dataset.

The Actor is using OCR Space service to perform the OCR.

Pricing

The Actor is using Pay-per-event pricing mode. There is flat fee for spinning the Actor up, and then for every successfully processed document page.

Documents failed to be processed are not counted.

For actual prices, see Actor pricing info.

Output

Each processed url is stored in default dataset in following shape:

{
    text: string, // The full text extracted from document.
    language: string, // Language of the document.
    url: string, // Original url.
    raw: object, // Raw data from the underlying OCR service. May change in the future.
}

In the processing ended up with error, the shape is different:

{
    url: string, // Original url.
    error: string, // Error message.
}

Planned features

The Actor is Work-in-progress - stay tuned for new features.

Integrability improvements

• Process all images and PDF files in provided key value store
• Process all urls from given dataset.