doc-redaction-app

To perform initial PII redaction on PDFs and images using a Gradio API, you can orchestrate calls to the /doc_redact or /redact_document endpoints, applying first-pass workflows with selectable OCR methods and PII entity lists.

You download redacted PDF artifacts from a Gradio server using deterministic download handling that fetches generated outputs via URL-encoded server paths from the gradio_api/file URL, optionally including HF authorization headers.

The difference between /doc_redact and /redact_document endpoints lies in their endpoint-specific parameter requirements for OCR and PII entity selection, allowing flexible API orchestration for different document redaction deployments.

Yes, you can choose specific OCR methods and PII entity lists for your document redaction workflow, enabling customized first-pass PII redaction tailored to your deployment requirements and data types.

After downloading the initial redacted documents and review artifacts, you should hand off the deliverables to later review and apply passes or parallel page-level review workflows for downstream processing.

Downloaded redaction artifacts fail to validate as non-empty deliverables when the Gradio API orchestration encounters endpoint parameter mismatches or incorrect server path URL-encoding during the redaction workflow execution.