{ "guide": { "slug": "philter-ucsf", "title": "Philter UCSF Guide", "subtitle": "A source-qualified guide to when Philter UCSF fits the locally deployable text redaction job.", "summary": "Philter UCSF is eligible for the clinical-note subset of local text de-identification; a specialized alternative, not a general-language drop-in.", "category": { "slug": "software", "name": "Software", "description": "Fact sheets for developer tools, AI systems, software platforms, and technical buying or adoption decisions." }, "tags": [ "software", "privacy", "pii", "redaction", "philter-ucsf" ], "version": "2026-10-18.1", "publishedAt": "2026-10-18", "r2JsonKey": "guides/philter-ucsf/2026-10-18.1.json", "updatedAt": "2026-10-18", "vaultPath": "Research/Technology/Software Development/AI Tools/Philter UCSF Guide.md", "freshness": "Checked 2026-10-18 against the cited sources. Official product claims, independent review/community signals, and evidence limits are reported; no hands-on benchmark, installation test, or privacy audit is claimed.", "heroLabel": "Locally deployable text redaction", "archetype": "comparison", "thesisCards": [ { "label": "Verdict", "title": "Choose it for clinical-note filtering.", "body": "Philter UCSF fits the clinical-note subset of local text de-identification. It is a specialized alternative rather than a general-language drop-in." }, { "label": "Workflow", "title": "Remove protected health information from clinical text.", "body": "Configure the filters and input/output paths for clinical plain-text notes, then inspect the redacted output. Evaluate it on representative clinical text before extending the workflow to another domain." }, { "label": "Boundary", "title": "Check the limit before committing", "body": "The key constraint is Clinical-domain focus; general chat/email quality is not established. No source here proves a universal performance, privacy, or reliability outcome." } ], "stats": [], "sections": [ { "id": "decision", "eyebrow": "Decision", "title": "Choose Philter UCSF when its documented boundary matches the job.", "body": "Philter UCSF fits the clinical-note subset of local text de-identification. The cited sources describe it as a locally deployable redaction option. Choose it when clinical-note de-identification is the job. Its clinical-domain focus does not establish quality for general chat or email." }, { "id": "setup", "eyebrow": "Setup", "title": "Start by confirming deployment and the data path.", "body": "Confirm this setup detail before you build a routine: Local Python command-line package. Then verify the processing boundary — Configurable filtering pipeline with optional evaluation annotations — and keep the first sample low-risk until the account, device, or runtime behavior matches the source description." }, { "id": "workflow", "eyebrow": "Workflow", "title": "Workflow: remove protected health information from clinical plain text.", "body": "Use Philter UCSF to remove protected health information from clinical plain text. Customization is described as filter configuration and input/output paths. Keep the workflow small enough to review its output and control settings; the cited sources do not establish a universal quality or speed result." }, { "id": "comparison", "eyebrow": "Alternatives", "title": "Compare Philter UCSF by the constraint that decides the work.", "body": "The five-guide table retains the current guide and four siblings in the locally deployable text redaction group. Values come from the cited source cells; an Unknown cell means the sources did not support a stronger statement.", "table": { "columns": [ "Guide", "Workflow", "Deployment", "Input", "Entry cost", "Method", "Customization", "Limitation" ], "rows": [ [ "OpenAI Privacy Filter", "Model-based text redaction with a local CLI and Python API.", "Local CPU or GPU; checkpoints downloaded if absent.", "Text strings, files and pipes.", "Open repository; compute and operational costs remain.", "Token classification with span decoding.", "Runtime precision/recall presets, evaluation and fine-tuning commands.", "Validate detection on the target corpus; community scores change with matching criteria." ], [ "Presidio", "PII identification followed by configurable anonymization.", "Python module or self-hosted HTTP service/container.", "Text and detected entity spans.", "Open-source deployment; self-hosting/integration cost separate.", "Analyzer plus separate anonymizer operators.", "Replace, mask, redact or encrypt; reversible operators need corresponding metadata and keys.", "Anonymizer acts on supplied detections; missing detections are not repaired by replacement." ], [ "scrubadub", "Replace personally identifiable spans in free text.", "Python package invoked inside local scripts.", "Free-text strings.", "Apache-2.0 software; dependency and operation costs separate.", "Detectors identify spans; postprocessors control replacement.", "Optional address, spaCy and Stanford detector packages; locale configuration.", "Extra detector dependencies are needed for some entity types." ], [ "Limina on-premises", "Detect and de-identify sensitive data; this comparison covers text.", "Container in customer on-premises environment or VPC.", "Text; product also supports documents, images and audio.", "Enterprise custom pricing; free hosted Starter is not the local-deployment price.", "Commercial entity detection and de-identification engine.", "Enterprise model-version pinning and guided integration/tuning.", "Local availability and commercial terms require vendor agreement." ], [ "Philter UCSF", "Remove protected health information from clinical plain text.", "Local Python command-line package.", "Plain-text file or directory of clinical notes.", "BSD-3-Clause software; integration and compute separate.", "Configurable filtering pipeline with optional evaluation annotations.", "Filter configuration and input/output paths.", "Clinical-domain focus; general chat/email quality is not established." ] ] } }, { "id": "review-signal", "eyebrow": "Review signal", "title": "Independent context for Philter UCSF is bounded.", "body": "UCSF authors / JAMIA Open (A certified de-identification system for all clinical text documents) reports: The available paper summary discusses the clinical system’s development and precision concerns. Limit: No numeric result from this paper was independently verified for this guide. These signals add context, but they do not establish comparative performance.", "bullets": [ "UCSF authors / JAMIA Open: The available paper summary discusses the clinical system’s development and precision concerns. (No numeric result from this paper was independently verified for this guide.)" ] }, { "id": "caveats", "eyebrow": "Caveats", "title": "Keep the source boundary visible while you decide.", "body": "The current cited cost statement is: BSD-3-Clause software; integration and compute separate. Privacy or processing control is described as: Privacy controls are not fully described in the cited sources. The principal product boundary is: Clinical-domain focus; general chat/email quality is not established. These are source-qualified statements, not an installation result, benchmark, security audit, or retention guarantee.", "bullets": [ "The paper summary provides clinical context, but no independently verified numeric result supports a performance comparison here.", "Repository and published system versions may differ. No claimed compliance certification or accuracy number." ] }, { "id": "decision-surface", "eyebrow": "First check", "title": "Make the first adoption check specific to Philter UCSF.", "body": "Run a labeled sample from the target corpus through detection and replacement, then inspect misses, over-redaction, and logs before sending real data through the pipeline." } ], "sources": [ { "id": "opf-repo", "title": "Privacy Filter repository and CLI", "publisher": "OpenAI", "url": "https://github.com/openai/privacy-filter", "sourceType": "official", "accessedAt": "2026-10-18" }, { "id": "presidio-home", "title": "Presidio overview", "publisher": "Data Privacy Stack", "url": "https://presidio.dataprivacystack.org/", "sourceType": "official", "accessedAt": "2026-10-18" }, { "id": "presidio-anonymizer", "title": "Presidio Anonymizer", "publisher": "Data Privacy Stack", "url": "https://presidio.dataprivacystack.org/anonymizer/", "sourceType": "official", "accessedAt": "2026-10-18" }, { "id": "scrubadub-repo", "title": "scrubadub README and license", "publisher": "LeapBeyond", "url": "https://github.com/LeapBeyond/scrubadub", "sourceType": "official", "accessedAt": "2026-10-18" }, { "id": "scrubadub-docs", "title": "scrubadub stable documentation", "publisher": "scrubadub maintainers", "url": "https://scrubadub.readthedocs.io/en/stable/", "sourceType": "official", "accessedAt": "2026-10-18" }, { "id": "limina-plans", "title": "Pricing and deployment FAQ", "publisher": "Limina", "url": "https://www.getlimina.ai/en/pricing", "sourceType": "official", "accessedAt": "2026-10-18" }, { "id": "philter-repo", "title": "Philter UCSF README", "publisher": "UCSF / BCHSI", "url": "https://github.com/BCHSI/philter-ucsf", "sourceType": "official", "accessedAt": "2026-10-18" }, { "id": "opf-community-eval", "title": "GLiNER large-v2.1 vs Privacy Filter case study", "publisher": "Gaurav Vij / GitHub", "url": "https://github.com/gauravvij/pii-filter-model-eval", "sourceType": "community", "accessedAt": "2026-10-18" }, { "id": "philter-paper", "title": "A certified de-identification system for all clinical text documents", "publisher": "UCSF authors / JAMIA Open", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC10320112/", "sourceType": "review", "accessedAt": "2026-10-18" } ], "claims": [ { "id": "research-philter-ucsf-philter-ucsf-workflow", "claim": "Remove protected health information from clinical plain text.", "sourceId": "philter-repo", "evidenceType": "official", "confidence": "high", "caveat": "Source-qualified product description; no independent performance test was run.", "guideLocation": "Workflow: remove protected health information from clinical plain text.", "sectionId": "workflow" }, { "id": "research-philter-ucsf-philter-ucsf-deployment", "claim": "Local Python command-line package.", "sourceId": "philter-repo", "evidenceType": "official", "confidence": "high", "caveat": "Source-qualified product description; no independent performance test was run.", "guideLocation": "Start by confirming deployment and the data path.", "sectionId": "setup" }, { "id": "research-philter-ucsf-philter-ucsf-input", "claim": "Plain-text file or directory of clinical notes.", "sourceId": "philter-repo", "evidenceType": "official", "confidence": "high", "caveat": "Source-qualified product description; no independent performance test was run.", "guideLocation": "Start by confirming deployment and the data path.", "sectionId": "setup" }, { "id": "research-philter-ucsf-philter-ucsf-entryCost", "claim": "BSD-3-Clause software; integration and compute separate.", "sourceId": "philter-repo", "evidenceType": "official", "confidence": "high", "caveat": "Observed source wording; regional pricing, billing cadence, and future plan changes are not inferred.", "guideLocation": "Choose Philter UCSF when its documented boundary matches the job.", "sectionId": "decision" }, { "id": "research-philter-ucsf-philter-ucsf-method", "claim": "Configurable filtering pipeline with optional evaluation annotations.", "sourceId": "philter-repo", "evidenceType": "official", "confidence": "high", "caveat": "Source-qualified product description; no independent performance test was run.", "guideLocation": "Workflow: remove protected health information from clinical plain text.", "sectionId": "workflow" }, { "id": "research-philter-ucsf-philter-ucsf-customization", "claim": "Filter configuration and input/output paths.", "sourceId": "philter-repo", "evidenceType": "official", "confidence": "high", "caveat": "Source-qualified product description; no independent performance test was run.", "guideLocation": "Workflow: remove protected health information from clinical plain text.", "sectionId": "workflow" }, { "id": "research-philter-ucsf-philter-ucsf-limitation", "claim": "Clinical-domain focus; general chat/email quality is not established.", "sourceId": "philter-repo", "evidenceType": "inference", "confidence": "medium", "caveat": "This is a scope inference; published clinical results do not establish quality on general chat or email.", "guideLocation": "Keep the source boundary visible while you decide.", "sectionId": "caveats" }, { "id": "research-philter-ucsf-signal-philter-paper", "claim": "UCSF authors / JAMIA Open reports: The available paper summary discusses the clinical system’s development and precision concerns.", "sourceId": "philter-paper", "evidenceType": "review", "confidence": "medium", "caveat": "No numeric result from this paper was independently verified for this guide.", "guideLocation": "Independent context for Philter UCSF is bounded.", "sectionId": "review-signal" } ], "relatedGuides": [ { "slug": "openai-privacy-filter", "role": "alternative", "rank": 1, "reason": "The existing locally runnable model/CLI baseline for contextual text redaction." }, { "slug": "presidio-text-redaction", "role": "alternative", "rank": 2, "reason": "A local integration toolkit when the evaluator needs configurable recognizers and multiple replacement operators." }, { "slug": "scrubadub-text-redaction", "role": "alternative", "rank": 3, "reason": "A locally embedded Python option for relatively simple text-cleaning pipelines with configurable detectors." }, { "slug": "limina-local-redaction", "role": "alternative", "rank": 4, "reason": "The commercial local-deployment option for organizations that want vendor-supported de-identification." } ], "comparison": { "title": "Locally deployable text redaction alternatives", "intro": "This table keeps philter-ucsf beside four listed alternatives using source-qualified workflow, platform, processing, cost, control, and limitation notes. It is an editorial fit map, not a performance test or market ranking.", "alternatives": [ "openai-privacy-filter", "presidio-text-redaction", "scrubadub-text-redaction", "limina-local-redaction" ], "includeCurrent": true, "dimensions": [ { "id": "workflow", "label": "workflow", "values": { "openai-privacy-filter": "Model-based text redaction with a local CLI and Python API.", "presidio-text-redaction": "PII identification followed by configurable anonymization.", "scrubadub-text-redaction": "Replace personally identifiable spans in free text.", "limina-local-redaction": "Detect and de-identify sensitive data; this comparison covers text.", "philter-ucsf": "Remove protected health information from clinical plain text." } }, { "id": "deployment", "label": "deployment", "values": { "openai-privacy-filter": "Local CPU or GPU; checkpoints downloaded if absent.", "presidio-text-redaction": "Python module or self-hosted HTTP service/container.", "scrubadub-text-redaction": "Python package invoked inside local scripts.", "limina-local-redaction": "Container in customer on-premises environment or VPC.", "philter-ucsf": "Local Python command-line package." } }, { "id": "input", "label": "input", "values": { "openai-privacy-filter": "Text strings, files and pipes.", "presidio-text-redaction": "Text and detected entity spans.", "scrubadub-text-redaction": "Free-text strings.", "limina-local-redaction": "Text; product also supports documents, images and audio.", "philter-ucsf": "Plain-text file or directory of clinical notes." } }, { "id": "entry-cost", "label": "entry cost", "values": { "openai-privacy-filter": "Open repository; compute and operational costs remain.", "presidio-text-redaction": "Open-source deployment; self-hosting/integration cost separate.", "scrubadub-text-redaction": "Apache-2.0 software; dependency and operation costs separate.", "limina-local-redaction": "Enterprise custom pricing; free hosted Starter is not the local-deployment price.", "philter-ucsf": "BSD-3-Clause software; integration and compute separate." } }, { "id": "method", "label": "method", "values": { "openai-privacy-filter": "Token classification with span decoding.", "presidio-text-redaction": "Analyzer plus separate anonymizer operators.", "scrubadub-text-redaction": "Detectors identify spans; postprocessors control replacement.", "limina-local-redaction": "Commercial entity detection and de-identification engine.", "philter-ucsf": "Configurable filtering pipeline with optional evaluation annotations." } }, { "id": "customization", "label": "customization", "values": { "openai-privacy-filter": "Runtime precision/recall presets, evaluation and fine-tuning commands.", "presidio-text-redaction": "Replace, mask, redact or encrypt; reversible operators need corresponding metadata and keys.", "scrubadub-text-redaction": "Optional address, spaCy and Stanford detector packages; locale configuration.", "limina-local-redaction": "Enterprise model-version pinning and guided integration/tuning.", "philter-ucsf": "Filter configuration and input/output paths." } }, { "id": "limitation", "label": "limitation", "values": { "openai-privacy-filter": "Validate detection on the target corpus; community scores change with matching criteria.", "presidio-text-redaction": "Anonymizer acts on supplied detections; missing detections are not repaired by replacement.", "scrubadub-text-redaction": "Extra detector dependencies are needed for some entity types.", "limina-local-redaction": "Local availability and commercial terms require vendor agreement.", "philter-ucsf": "Clinical-domain focus; general chat/email quality is not established." } } ] }, "revisions": [ { "version": "2026-10-18.1", "date": "2026-10-18", "note": "Initial dedicated guide from the 2026-10-18 cited sources." } ] }, "publicUrl": "https://fieldguides.co/g/philter-ucsf" }