Rules

Rules are workspace-scoped decisioning logic. They evaluate JSON Logic expressions against the full decisioning context — across any model and relationship within that context. Rules are not tied to a single model.

Endpoints

Method	Path	Description
POST	/rules	Create a rule
GET	/rules	List rules
GET	/rules/{ruleId}	Get a rule
PUT	/rules/{ruleId}	Update a rule
DELETE	/rules/{ruleId}	Delete a rule
PUT	/rules/{ruleId}/status/{targetStatus}	Lifecycle transition

Lifecycle

Rules have their own lifecycle, independent of models and contexts:

draft → active → archived

Status	Meaning
draft	The rule is being configured. It does not evaluate.
active	The rule is live. It evaluates against all relevant contexts.
archived	The rule is retired. Active evaluations are removed; the definition is preserved.

Create a rule

POST /rules

Request

{
  "name": "LTV Threshold Check",
  "description": "Applicants must have a loan-to-value ratio below 80%",
  "category": "affordability",
  "severity": "hard",
  "expression": {
    "<": [{ "var": "property.ltv_ratio" }, 80]
  }
}

Field	Required	Description
name	Yes	Human-readable rule name
description	No	What the rule checks
category	No	Freeform grouping string (e.g., affordability, identity, compliance)
severity	Yes	Impact level: hard, soft, or info
expression	Yes	Valid JSON Logic expression

Severity levels

Severity	Meaning
hard	Blocking. The decision cannot proceed if this rule fails.
soft	Warning. Flags an issue but does not block the decision.
info	Informational. Logged for audit but triggers no action.

Response 201 Created

{
  "ruleId": "01957c3a-abcd-7000-8000-000000000001",
  "name": "LTV Threshold Check",
  "description": "Applicants must have a loan-to-value ratio below 80%",
  "category": "affordability",
  "severity": "hard",
  "expression": {
    "<": [{ "var": "property.ltv_ratio" }, 80]
  },
  "inputPaths": ["property.ltv_ratio"],
  "status": "draft",
  "createdBy": "user_01957c3a-...",
  "createdAt": "2026-02-20T10:10:00Z",
  "updatedAt": "2026-02-20T10:10:00Z"
}

The server validates the JSON Logic expression and resolves all var references against existing model fields in the workspace. Invalid expressions or unresolvable paths return 422 Unprocessable Entity.

inputPaths is derived from the expression at creation time — it is a read-only field listing all field paths the rule depends on.

Get a rule

GET /rules/{ruleId}

Response 200 OK

Same shape as the create response.

List rules

GET /rules

Supports standard filtering and sorting.

Useful filters

GET /rules?filter[status][eq]=active              # Active rules only
GET /rules?filter[category][eq]=affordability     # By category
GET /rules?filter[severity][eq]=hard              # Hard blocks only
GET /rules?sortBy=updatedAt:desc                  # Most recently modified

Update a rule

PUT /rules/{ruleId}

Rule updates use PUT (full replacement) rather than PATCH (partial update). Provide the complete rule definition in every update request — any omitted fields will be cleared. This differs from other resources in the API (models, records, contexts) which support partial updates via PATCH.

Updating an active rule bumps the version and marks all affected evaluation rows stale for recomputation.

Request

{
  "name": "LTV Threshold Check",
  "description": "Updated: LTV must be below 85%",
  "expression": {
    "<": [{ "var": "property.ltv_ratio" }, 85]
  }
}

Response 200 OK

Returns the full updated rule with the new version.

Lifecycle transition

PUT /rules/{ruleId}/status/{targetStatus}

Effects

• Activating a rule creates evaluation rows for all relevant contexts and triggers initial evaluation.

• Archiving a rule removes active evaluations and preserves the definition.

Error cases

Scenario	Status
Invalid transition (e.g., archived → draft)	422 Unprocessable Entity

Delete a rule

DELETE /rules/{ruleId}

Performs a soft delete — the rule is marked as deleted and excluded from system operations but retained for auditing.

Response 204 No Content