Call a Specialist Today! (02) 9388 1741
Govern AI Output with Real-Time Guardrails
Protect users, data, and your brand with automated input/output filtering PII detection, content moderation, topic restrictions, and custom logic all enforced at the gateway.
Stop Harmful Content Before It Reaches Users
Guardrails sit in the gateway layer applied consistently to every request, regardless of which model, framework, or team is making the call.
PII Detection & Redaction
Automatically detect and block names, emails, credit card numbers, and other PII in prompts and responses before they reach external providers or end users.
- 20+ PII entity types
- Prompt and response scanning
- Auto-redaction or blocking
- Full event log
Content Moderation
Apply topic restrictions, toxicity filters, and hate-speech detection at the gateway level no model fine-tuning required.
- Topic restriction rules
- Toxicity and hate-speech filters
- Configurable thresholds
- Works across all models
Regex & Custom Rules
Define custom regex patterns or logic to match proprietary data formats, internal keywords, or domain-specific risk patterns.
- Custom regex patterns
- Keyword blocklists
- Domain-specific rules
- Per-team configuration
Third-Party Safety Integrations
Connect Portkey guardrails to Aporia, Pillar, Patronus AI, and other third-party safety providers via a unified interface.
- Aporia, Pillar, Patronus AI
- Unified guardrail interface
- Mix built-in and partner checks
- Pluggable architecture
Input & Output Checks
Run guardrails on both the prompt going in and the model response coming out for end-to-end protection on every interaction.
- Pre-request input scanning
- Post-response output scanning
- Block or redact on trigger
- Full audit trail
Real-Time Alerts & Logging
Get notified when guardrails fire. Every blocked event logged with full context for compliance reporting and incident review.
- Instant trigger alerts
- Full blocked-event log
- Exportable for compliance
- Per-rule analytics
Protection at the Infrastructure Layer Not the Application Layer
Bolting safety onto individual applications is fragile and inconsistent. Portkey Guardrails enforce policies at the gateway so every team and every model is covered automatically.
- PII protection before data leaves your environment
- Content moderation without model fine-tuning
- Consistent enforcement across all teams and models
- Full audit log for compliance and security review