Offline-first pipeline that turns intake PDFs and CSVs into deduplicated nonprofit constituent records: OCR extraction, probabilistic matching via Splink, a WCAG 2.2 AA human review queue, consent-aware write-back to CiviCRM, Salesforce, and webhooks, a VAWA/FVPSA DV privacy pack, Docker self-host, and a tamper-evident provenance log.
python docker pdf privacy ocr offline-first accessibility entity-resolution nonprofit self-hosted wcag salesforce civic-tech human-services deduplication civicrm domestic-violence data-provenance data-intake splink
-
Updated
Jul 14, 2026 - Python