Gini API documentation


Gini provides an information extraction system for analyzing documents (e. g. invoices or contracts), specifically information such as the document sender or the amount to pay in an invoice.

The API supports documents in the formats PDF, GIF, JPEG, PNG, and TIFF.

Furthermore there is the distinction between native PDF files (i. e. files digitally created from Microsoft Word documents and containing textual information) and scanned PDF files, created by scanning paper documents.

After a file has been submitted and validated, Gini first extracts the textual contents of the document. Native PDF files already contain this information and are analyzed accordingly. Scanned documents (like the aforementioned) image files do not provide easily processable information directly, so Gini has to preprocess these files using Optical Character Recognition (OCR) and other methods to obtain their content.

Once the layout and textual information is available for the uploaded document, Gini starts extracting semantic and meta information about the document, such as the document sender (e. g. name, address) or the document type (e. g. invoice, contract).

In the unlikely case that Gini was unable to infer the information correctly, it is possible to correct the extractions manually (e. g. by selecting the correct amount to pay on the document) and submit it back to the API. This also helps to improve the learning extraction system over time.

If you have any questions about the Gini API and the functionality it provides, please contact us via