CyberCage Docs

Appearance

Network Inspection ​

Monitor and control AI traffic with PII detection, content rules, and provider management.

Accessing Network Inspection ​

Navigate to Guardrails → Network Inspection in the Dashboard.

Tabs ​

AI Providers ​

Configure which AI providers to monitor.

Features:

  • Enable/disable inspection per provider
  • Block traffic to specific providers
  • Block file uploads
  • Set request size limits
  • View provider domains

Supported Providers: OpenAI, Anthropic, Google AI, Azure OpenAI, Cohere, Mistral, and more.

Violations ​

View detected policy violations in real-time.

Filter by:

  • Violation type (PII, content rule, secret)
  • Severity (critical, high, medium, low)
  • Provider
  • Date range

Each violation shows:

  • Timestamp
  • Provider and domain
  • Violation type and severity
  • Action taken (logged, blocked, redacted)
  • Device and user info

Content Rules ​

Create rules to detect sensitive content patterns.

Rule Types:

  • Keyword - Match specific words or phrases
  • Regex - Match patterns with regular expressions

Actions:

  • Log - Record but allow
  • Block - Reject the request
  • Redact - Replace matched content

Built-in Templates:

  • API keys (GitHub, OpenAI, AWS, Stripe, Slack)
  • Private keys (RSA, EC, OpenSSH, PGP)
  • Database connection strings
  • Cloud provider tokens

PII Detection ​

Configure which PII types to detect and how to handle them.

Entity Categories:

  • Personal - Names, emails, phone numbers, addresses
  • Financial - Credit cards, IBANs, crypto addresses
  • Network - IP addresses, URLs
  • Identifiers - SSNs, passport numbers, driver licenses

Per-entity settings:

  • Enable/disable detection
  • Set action (log, block, redact)
  • Set severity level

Settings ​

Toggle inspection features at the organization level.

Available Settings:

  • Network inspection enabled
  • PII detection enabled
  • Content rules enabled
  • Image PII detection enabled
  • Request/response logging

Next Steps ​