TrigGuard
TRIGGUARD INTEGRATION
Open Console

Integration architecture

Put TrigGuard between intent and execution

Integrate through the SDK, execution gateway, or an execution-specific adapter. TrigGuard evaluates the requested action before it reaches the system that can execute it.

System architecture → /architecture. How to implement → Quickstart.

  1. Agent / Application / Automation
  2. Integration point · SDK / API / Adapter
  3. TrigGuard · Authorization gate
  4. PERMIT / DENY / ESCALATE / SILENCE
  5. Tool / Cloud / DB / Robot
  6. Receipt

Choose your integration point

Three ways to sit on the execution path

  • Inline SDK

    Application and agent runtimes

    Lowest integration distance

  • Gateway

    APIs, CI/CD, cloud, shared infrastructure

    Central authorization boundary

  • Adapter

    Robotics, OT, execution-specific systems

    Bind authorization to the execution interface

Same boundary, different surfaces

Authorization model stays constant

The gate is always before the irreversible action. Only the surface changes.

Surface Before Authorized execution
AI agentstool calltool / API
CI/CDreleasedeployment
Cloud / IaCapplyinfrastructure
Databasemutationdatabase
Roboticscommandcontroller / actuator
Enterpriseworkflow actiontarget system

One example

Authorize, then execute

Agent
  ↓ proposes
deploy.release
  ↓
authorize({
  surface: "deploy.release",
  actor: "release-agent",
  target: "production"
})
  ↓
TrigGuard
  ↓
PERMIT + receipt
  ↓
Deployment executes

See implementation → Quickstart Request schema → API reference Decision semantics → Protocol

Ready to integrate?

Start with the Quickstart, or talk to us about a production architecture.

Start building Talk to us