Agent SDK

Build custom agents. Upgrade-safe.

The slx.* extension namespace keeps your custom agents isolated from core SyteLine. Build, test, and deploy without risk.

Policy-enforced
slx.* namespace
Hot deploy

Interactive Demo

Build an agent — step by step.

Walk through the agent creation wizard and see the generated code.

Workflow

Five steps to your first agent

01

Install the SDK

Add the Agent SDK to your project with a single command.

npm install @syteray/agent-sdk
02

Scaffold an agent

Generate the boilerplate for your custom agent in the slx.* namespace.

sl agent scaffold slx.demand_forecaster --type analytics
03

Configure policies

Define what your agent can access. Policy engine enforces boundaries.

// slx.demand_forecaster/policy.yaml
permissions:
  read: [SLItems, SLInventory, SLJobs]
  write: []
  invoke: [SLForecastMethods]
limits:
  max_rows: 10000
  timeout: 30s
04

Build the logic

Write your agent using TypeScript. Access IDOs, run queries, and generate insights.

import { Agent, IDO, Report } from '@syteray/agent-sdk';

export default class DemandForecaster extends Agent {
  name = 'slx.demand_forecaster';
  version = '1.0.0';

  async execute(context) {
    const items = await IDO.query('SLItems', {
      filter: { status: 'Active' },
      limit: 1000
    });

    const forecast = this.analyze(items);
    return Report.create({
      title: 'Demand Forecast',
      data: forecast,
      confidence: 0.94
    });
  }
}
05

Deploy

Deploy your agent to your on-prem instance. Air-gapped and secure.

sl agent deploy slx.demand_forecaster --env production
# Agent live on :8091 — air-gapped ✓
# Policy validated ✓
# Namespace: slx.demand_forecaster — upgrade-safe ✓

Build your first agent today

150+ built-in agents ship free. Add your own with the SDK and deploy them in minutes, not weeks.