Custom trained extractors in UiPath Document Understanding are trained directly within the Document Understanding application hosted in UiPath Automation Cloud, using features such as Data Extraction Training and One-Click Extraction. These tools let users label sample documents and train ML extraction models without manually configuring Datasets, Pipelines, or ML Skills in AI Center. Studio Desktop and Studio Web are used later to integrate the trained extractor into an automation workflow (e.g., via the DocumentUnderstanding.Activities package), while Orchestrator handles deployment and execution, not model training.
Community Discussion