Documentation Index

Fetch the complete documentation index at: https://docs.aifabrix.ai/llms.txt

Use this file to discover all available pages before exploring further.

Entity basics, types, and resource type

Prev Next

Set entityType and root resourceType on each Business Entity manifest so sync, storage, RBAC, and AI exposure follow the correct contract.

Prerequisites

  • Connected System type chosen (Basics and kind)
  • One JSON file per Business Entity under integration/<systemKey>/
  • Builder CLI authenticated — aifabrix auth status green

Where it lives

<datasourceKey>.json → top-level entityType, resourceType, key, displayName, systemKey.

Manifest sections: datasource.basics, datasource.entityType, datasource.resourceType. Both identity fields sit at the datasource root — never nest resourceType under fieldMappings.

Builder MCP

Manifest sections datasource.basics, datasource.entityType, datasource.resourceType
UI tab Overview / Technical readiness
Help topics section, goldenExamples, goldenExample
Repair hooks fieldMappings, metadata schema alignment

Entity type recipes: Business Entities guides.

How to set

  1. Create or open the entity file — scaffold the Connected System folder first:
aifabrix create <systemKey> --type external
aifabrix repair <systemKey> --dry-run
  1. Classify each Business Entity — what it primarily stores or represents.
  2. Set entityType per schema enum:
entityType Use when Enterprise Knowledge role
recordStorage CRM-style entities (customers, deals, cases) Structured records with metadata
documentStorage Files with metadata, no vector index Document catalog and search
vectorStore Embeddings and semantic search Vector-backed retrieval
static Reference data with infrequent change Stable lookup tables
service Callable capability without record storage Governed operations
  1. Align fieldMappings and sync — record types use business entities; document types emphasize file metadata.
  2. List catalog keys — reuse standard tokens when possible:
aifabrix resource-type list
aifabrix resource-type get --key customer
  1. Create missing catalog keys before first upload:
aifabrix resource-type create --key customer --display-name "Customer" --category business
  1. Set resourceType at the datasource root and validate:
aifabrix datasource validate <datasourceKey>
aifabrix validate <systemKey>
  1. Continue in entity guidesrecord storage, document storage, or other guides overview for schema bands beyond identity.

Defaults and examples

Element Purpose Example
entityType Storage and sync model recordStorage, documentStorage
resourceType Business catalog token customer, document, deal
key Entity identifier example-crm-customers
systemKey Parent Connected System Must match system JSON

Example identity fragment:

{
  "key": "example-crm-customers",
  "displayName": "Customers",
  "systemKey": "example-crm",
  "entityType": "recordStorage",
  "resourceType": "customer",
  "primaryKey": ["externalId"]
}

Illustrative classification goals:

  • CRM companies → recordStorage + customer
  • SharePoint library → documentStorage + document
  • reference lookup table → static + catalog key matching business vocabulary

Validate

aifabrix repair <systemKey> --dry-run
aifabrix datasource validate <datasourceKey>
aifabrix validate <systemKey>

Wrong entityType often surfaces as sync or CIP compile errors — not auth failures. Re-run validate after changing resourceType; protection and subscription filters reference catalog tokens at runtime.

Common mistakes

Mistake Fix
resourceType under fieldMappings Move to root next to entityType
Invented catalog token Create with resource-type create first
Document entity with recordStorage Switch entityType to match storage model
Skipping repair after identity edits repair syncs metadata schema hints

Limits

entityType drives sync bands, CIP templates, and UI tabs — changing it after publish may require re-modeling field mappings and protection. resourceType must exist in the platform catalog before upload; this page shows shape, not every catalog category. Complex multi-entity vendors still use one Business Entity file per governed resource — do not merge unrelated resources into a single JSON file.