freight-template-learner

Template learning automatically extracts BOL and POD fields by capturing unique document structures per client. You process broker documents using existing client templates, eliminating the need to code new parsers for each format.

Template learning is a training-mode mechanism that learns broker-specific document formats by processing corrections. It stores per-client mappings locally to automatically extract BOL, POD, and related fields across multiple brokers.

Yes, data extraction accuracy improves through a feedback loop of broker corrections. The training-mode driven improvements capture corrections during extraction to continuously refine per-client templates for future document processing.

Template learning works with OCR or document text input and applies across multiple brokers without code changes. It requires local storage and stores per-client mappings in a local JSON file for quick deployment across different broker formats.

Yes, local storage is required to store per-client mappings in a local configuration file. This local storage enables the template learner to quickly deploy and apply learned broker-specific document formats during extraction.

Template learning handles multiple broker document formats by automatically learning per-client structures without coding. It captures unique BOL, POD, and related document structures, reducing maintenance and setup time across different brokers.