PDFScan

PDFScan represents the technological evolution in the automation of the accounts payable cycle for the Dolibarr ecosystem, overcoming the limitations of traditional OCR systems thanks to an advanced Artificial Intelligence engine. The module is designed to completely eliminate manual data entry for foreign supplier invoices, natively supporting documents from any international context (including complex markets such as the United States, China, and Russia) and accurately interpreting non-Latin alphabets, different currencies, and tax rates.
The operational flow is fully automated and optimized for maximum efficiency: the user uploads documents in PDF format—either individually or via bulk ZIP archives—and the system takes care of document management locally. An asynchronous process, orchestrated by Dolibarr's native cron every five years, sends the files to the ingestion service APIs, monitoring their progress through a precise status pipeline (pending, queued, processing, completed, error).
Once processing is complete, the algorithm not only extracts macroscopic data but also performs an analytical breakdown of the document line by line, faithfully recreating the structure of the details and tables. The end result is the automatic generation of a supplier invoice draft directly within the management system, complete with the original PDF attachment and ready for accounting validation. In the event of any registry anomalies or data inconsistencies, the system promptly traces the cause of the error in the checklist, ensuring complete process oversight.