Explicit PHI rules
Define the concrete tables and columns that must return a masked representation.
Healthcare workloads
Reduce PHI exposure on constrained read paths.
Configure the PostgreSQL columns that contain protected health information, route selected reads through Safe Boundary, and mask those values before they reach the caller.
Healthcare workloadsKeep configured PHI out of routine reads
PHI rule matchedDATABASE RESULTmember_id: 48291phone: 415-555-0182notes: diagnosis…
PHI boundaryExplicit column rules
PROTECTED RESULTmember: •••91phone: •••-•••-0182notes: [MASKED]
3 FIELDS
Why this exists
A focused safeguard, not a compliance shortcut
Healthcare data often sits in columns such as notes, member_id, phone, or address. Safe Boundary applies operator-defined schema and column rules; it does not inspect arbitrary values and autonomously discover PHI.
Read-only enforcement, dangerous-function denial, response masking, and local-first audit can support a broader HIPAA security program. Risk analysis, retention, access review, and contractual obligations remain organizational responsibilities.
What it does
Concrete controls, stated precisely.
Constrained agent reads
Use MCP read-only controls and result limits for AI workflows.
Decision audit
Retain allow, block, and mask outcomes in local formats for your evidence pipeline.
Control versus program
Use the product control as one panel in the compliance design.
| Decision | Safe Boundary can provide | Your healthcare program still needs |
|---|---|---|
| Data path | Configured masks, read-only enforcement, routed audit | Inventory, access review, retention, incident response |
| PHI scope | Explicit operator-defined columns | Organizational classification and clinical-text handling |
| Assurance | Evidence for decisions made at the boundary | Risk analysis, contracts, legal commitments, and independent assessment |
How it works
A visible enforcement sequence.
Classify
Use your data inventory to identify protected tables and columns.
Configure
Translate that inventory into explicit masking rules.
Route
Protect the approved application, REST, or MCP path.
Validate
Test masks, logging, retention, and bypass controls together.
Protected when
- ✓Configured PostgreSQL fields on routed paths
- ✓Unsafe SQL on enforcing paths
- ✓Policy-decision evidence for security operations
Important boundaries
- •No automatic PHI discovery or clinical-text classification
- •No claim that deployment alone makes a system HIPAA compliant
- •A BAA or certification must be confirmed separately as a legal commitment
Continue the story
Related Safe Boundary pages
Schema-aware masking →
Configure which columns are protected and preserve compatible response types.
MCP server →
Give AI tools a bounded PostgreSQL query surface with local audit.
Security model →
Understand trust boundaries, credentials, identity, and bypass conditions.
Test the boundary on a real access path.
Start in observe mode and validate the policy against representative traffic before you enforce it. That gives the team evidence for each decision, along with explicit bypass controls for the cases that need a deliberate exception.