Machine Layer

Structured access for automated systems.

The Machine Layer is the direct integration surface for automated systems, agents, and programmatic clients. All inputs follow structured intake standards. Human escalation paths are available only as downstream routing outputs — not as direct access.

View Endpoints Read Intake Standards

No Auth Required

Structured Input Only

Async Routing

API Endpoints

Available routes. All inputs are validated before processing.

POST/api/intakeSubmit a structured case for triage and routing.Auth: None

POST/api/classifyClassify a case type and return routing recommendation.Auth: None

GET/api/status/:idRetrieve the current status and routing outcome of a submitted case.Auth: Case Token

POST/api/agent/querySubmit a structured query to the front-end agent for automated resolution.Auth: None

GET/api/standardsRetrieve the current intake standards and access rules as structured data.Auth: None

Payload Schema

Required and optional fields for all submissions.

Field	Required	Values	Description
`type`	Required	`project \| partnership \| capital \| inquiry`	The classification of the submission.
`origin`	Required	`string`	Machine identifier or system name originating the request.
`payload`	Required	`object`	Structured case data. Schema varies by type. See /api/standards.
`priority`	Optional	`standard \| elevated`	Routing priority hint. Subject to system override.
`callback_url`	Optional	`string (URL)`	Webhook URL to receive status updates upon routing completion.

Response Codes

What each response means. Handle all of them.

200OK

Request received and queued for triage.

202Accepted

Request accepted and routing in progress.

400Bad Request

Payload is malformed or missing required fields.

403Forbidden

Access level insufficient for the requested path.

422Unprocessable

Payload is structurally valid but fails intake standards.

429Rate Limited

Too many requests. Apply exponential backoff.

Design Principles

How the Machine Layer is designed to behave.

These are the operating principles for automated clients. Build to them.

Stateless by defaultEvery request is self-contained. No session state is maintained. Idempotency is recommended for all submissions.

Structured inputs onlyUnstructured or free-text payloads are rejected. All inputs must conform to the intake schema for their type.

Triage before routingNo submission is routed without classification. Routing is a downstream output of triage, not an input assumption.

Callbacks are asyncStatus updates are delivered asynchronously via webhook. Polling /api/status is available but not preferred.

Example Request

A valid intake submission.

POST /api/intake
Content-Type: application/json

{
  "type": "project",
  "origin": "automated-client-v1",
  "priority": "standard",
  "callback_url": "https://your-system.example/webhook",
  "payload": {
    "title": "Infrastructure Resilience Assessment",
    "scope": "national",
    "urgency": "standard",
    "description": "Request for structured assessment of critical infrastructure...",
    "contact": {
      "entity": "Acme Infrastructure Group",
      "jurisdiction": "AU"
    }
  }
}

Ready to Integrate

Automated systems enter through the Machine Layer. Humans enter through the front-end.

All intake standards apply equally to machine and human submissions. Review the standards framework before building your integration.

Read Standards Talk to the Agent

REST API

JSON Schema

Webhook Callbacks