Agent Mask Reviews & Product Details

Agent Mask helps Snowflake teams de-identify sensitive text and documents without moving PII or PHI to an external redaction API. Install Agent Mask from Snowflake Marketplace, call one SQL function, and return de-identified output plus entity evidence directly inside Snowflake. It is built for the messy data teams actually need to use: support tickets, call transcripts, clinical notes, intake forms, chat logs, AI outputs, PDFs, scanned PDFs, images, DICOM files, DOCX, RTF, and ZIP archives. Because Agent Mask runs as a Snowflake Native App inside the customer’s Snowflake account, sensitive data stays where it already lives. There are no file transfers to Agent Mask, no external processing pipeline, and no separate API writeback workflow to govern. Run it wherever you already run SQL: worksheets, Airflow DAGs, dbt models, Dynamic Tables, stored procedures, notebooks, and scheduled tasks. Agent Mask is useful when redacted data still needs to support analytics, AI/ML, QA, support review, audits, or governed sharing. Teams can choose how each entity type is handled, including readable placeholders, fixed labels, character masks, deterministic hashes, synthetic replacements, reversible encryption, detect-only mode, and entity labels. It also supports custom entities, identity-aware consistency, and entity coreference resolution, so repeated references to the same person or organization can stay aligned across related text. Responses can include analyzer results, an entity ledger, detected values, replacements, document metadata, and redaction metadata for review. Use Agent Mask when your Snowflake warehouse contains valuable text or files that are too sensitive to hand directly to analysts, models, partners, or downstream systems.