Vector Store API

DRAFT — Internal Developer Use Only

This API reference is for internal development teams.

Overview

What it is: The vector store API manages collections of vector embeddings for semantic search and AI-powered queries. Collections store document embeddings that enable similarity search across identity data.

Endpoints

Method	Path	Description	Auth required
POST	/internal/v1/ai/vector_store/collection	Create a collection	JWT + API token
GET	/internal/v1/ai/vector_store/collection	List all collections	JWT + API token
GET	/internal/v1/ai/vector_store/collection/:id	Get collection details	JWT + API token
POST	/internal/v1/ai/vector_store/collection/:id/query	Query a collection	JWT + API token
DELETE	/internal/v1/ai/vector_store/collection/:id	Delete a collection	JWT + API token

---

POST /internal/v1/ai/vector_store/collection

Create a new vector store collection.

Request:

POST /internal/v1/ai/vector_store/collection
Authorization: Bearer <token>
Content-Type: application/json

{
  "name": "identity-embeddings",
  "description": "Vector embeddings for identity entity search",
  "dimensions": 1536
}

Response (201):

{
  "id": "collection-uuid-001",
  "name": "identity-embeddings",
  "status": "created",
  "documentCount": 0
}

---

POST /internal/v1/ai/vector_store/collection/:id/query

Execute a semantic query against a vector store collection.

Request:

POST /internal/v1/ai/vector_store/collection/collection-uuid-001/query
Authorization: Bearer <token>
Content-Type: application/json

{
  "query": "privileged accounts with access to production databases",
  "topK": 10,
  "threshold": 0.7
}

Response (200):

{
  "results": [
    {
      "id": "doc-001",
      "score": 0.95,
      "metadata": {
        "entityType": "account",
        "source": "PostgreSQL"
      },
      "content": "DBA admin account with full access to production database cluster"
    }
  ]
}