Selah integration guide

Selah is a pre-execution governance engine for AI agents. It decides what an agent is allowed to do and say before the agent acts, instead of logging what happened after. The model is simple: the agent proposes, the engine decides, and only a permitted action is executed.

What the engine enforces is a set of Agent Operating Procedures, or AOPs: clear, versioned rules for what an agent may do and say. You author them, run everything in shadow mode first to build a Shadow AOP Ledger, calibrate the rules against what the engine would do, and then switch to enforcement. This guide takes you from zero to production along that path.

How this guide is organized

Start with the foundation, then read the concept page, then follow the guide for your tenant type. The reference pages apply to all three.

Foundation, read first:

Getting started covers the core idea, the base URL, authentication, and your first call.
Agent Operating Procedures and the Shadow AOP Ledger defines the two terms the whole product is built on.

Pick your path, a full step by step from zero to production:

Enterprise, for a company running its own agents.
Platform, for a company whose product is agents serving many of their own customers.
Private, for a private or self-hosted agent deployment.

Reference, shared across all paths:

API reference, every endpoint with examples.
Webhooks and holds, the events and how a human steps in.
Going live, the shadow to enforcement checklist.
Glossary, every term defined once.

Before you start

You will need access to the Selah dashboard for your organization, an API key once you reach the integration step, and the agent code you want to govern.

A note on exact shapes: the request and response bodies in this guide are written to make the intent clear. The live API schema at https://api.selahcore.com/openapi.json is authoritative for exact field names, types, and status codes. Generate your client from it, and confirm any shape here against it before you ship.