Skip to main content

Interface: GuardrailConfig

Defined in: packages/agentos/src/core/guardrails/IGuardrailService.ts:212

Configuration for guardrail evaluation behavior.

Controls when and how often guardrails evaluate content. Use these settings to balance safety requirements against performance and cost constraints.

Example​

// Real-time PII redaction with rate limiting
const config: GuardrailConfig = {
  evaluateStreamingChunks: true,
  maxStreamingEvaluations: 50
};

// Cost-efficient final-only evaluation (default)
const config: GuardrailConfig = {
  evaluateStreamingChunks: false
};

Properties​

evaluateStreamingChunks?​

optional evaluateStreamingChunks: boolean

Defined in: packages/agentos/src/core/guardrails/IGuardrailService.ts:233

Enable real-time evaluation of streaming chunks.

When true, evaluates every TEXT_DELTA chunk during streaming. When false (default), only evaluates FINAL_RESPONSE chunks.

Performance Impact:

  • Streaming: Adds 1-500ms latency per TEXT_DELTA chunk
  • Final-only: Adds 1-500ms latency once per response

Cost Impact:

  • Streaming: May trigger LLM calls per chunk (expensive)
  • Final-only: Single evaluation per response (cost-efficient)

Use Cases:

  • Streaming (true): Real-time PII redaction, immediate blocking
  • Final-only (false): Policy checks needing full context, cost-sensitive

Default​

false

maxStreamingEvaluations?​

optional maxStreamingEvaluations: number

Defined in: packages/agentos/src/core/guardrails/IGuardrailService.ts:244

Maximum streaming evaluations per request.

Rate-limits streaming evaluations to control cost and performance. Only applies when evaluateStreamingChunks is true. After reaching the limit, remaining chunks pass through unevaluated.

Default​

undefined (no limit)