Docs

Quickstart

Connect an MCP client, save the sample sales order SO-03242607, then retrieve it and inspect its history in Neural Fabric.

Before you begin. You’ll need a Qorbit workspace, a provider API key for the model that powers Qora and classification, and an MCP client. The client connection is separate from the provider key. Step 1 walks through both.

1

Connect

Start a trial or sign in so Qorbit can provision Neural Fabric. In Settings → Models & voice, choose Anthropic / Claude, OpenAI, xAI / Grok, Moonshot AI / Kimi, or DeepSeek, enter its API key, select a model, test, and save. That provider and model power Qora and classification. The steps and key links are on Connect. Then connect one MCP client. Cloud assistants use OAuth. Editor and local clients use the bearer token shown only for those clients. That client authorization is not the provider API key. Review the permission prompt. Do not grant unrestricted access by default.

Sign in

Expected

You can open Neural Fabric and Connect. The Neural Fabric tools appear in the client. fabric_list with no required arguments returns entities, total, and showing.

If this stalls: wait for credentials to finish provisioning, then refresh. If tools are missing, confirm OAuth finished or that the bearer token matches Connect, then restart the local client.

2

Add the sample sales-order entity

This sample uses a sales order and quotation to demonstrate the API. The same entity, connection and history operations apply across supported business domains and information types.

Ask your connected client to call fabric_add with the sample arguments below.

fabric_add · sample 
{
  "entityType": "SalesOrder",
  "name": "SO-03242607",
  "properties": {
    "orderNumber": "SO-03242607",
    "customerName": "Clearview Commercial Lighting",
    "quoteReference": "QT-03242607",
    "customerPo": "CCL-2026-R117",
    "orderDate": "2026-04-02",
    "amount": 510105,
    "status": "open"
  },
  "actor": "docs-sample"
}

Expected

The tool saves the entity and returns an entityId. Copy that ID as YOUR_SALES_ORDER_ENTITY_ID. If you omitted entityId in the request, do not guess it.

If add fails: confirm the provider key is saved. Keep to 60 or fewer top-level property keys and string values under 3,500 characters.

3

Retrieve it

Search by order number, then load the full entity. fabric_search supports match (stored-entity lookup) and answer (a model-composed response). When mode is omitted, search defaults to match. Answer generation requires explicit mode: "answer". This identifier lookup omits mode and uses match.

fabric_search 
{
  "query": "SO-03242607",
  "entityType": "SalesOrder",
  "detail": true,
  "limit": 5
}
fabric_list 
{
  "entityId": "YOUR_SALES_ORDER_ENTITY_ID",
  "detail": true
}

Expected

Search returns the sales order in results, or results: [] if it is not in this workspace yet. List detail returns entity.id, entity.type, entity.name, entity.properties, and entity.deleted.

If search is empty: confirm fabric_add succeeded in this workspace, then retry with fabric_list and the returned ID.

4

Inspect it in Neural Fabric

Open Neural Fabric in the Qorbit app (/app/graph). Search for SO-03242607 or open /app/graph?entity=YOUR_SALES_ORDER_ENTITY_ID. The graph is a view of the same stored entity.

Select the sales order. Confirm customerName, quoteReference, and status. Then call:

fabric_provenance 
{
  "entityId": "YOUR_SALES_ORDER_ENTITY_ID"
}

Expected

History includes an entity_create entry. The actor field records the caller-supplied label docs-sample; it is not verified identity. The entry also includes timestamp and verification.status.

Next: see a derived connection.

Add the sample quotation below. Include the quotation identifier and customer name as entity properties. Those are supplied fields. Neural Fabric derives the connection to the sales order from shared values; you do not write the edge.

fabric_add · sample quotation 
{
  "entityType": "Quotation",
  "name": "QT-03242607",
  "properties": {
    "quoteNumber": "QT-03242607",
    "customerName": "Clearview Commercial Lighting"
  },
  "actor": "docs-sample"
}
fabric_graph 
{
  "mode": "root",
  "rootEntityId": "YOUR_SALES_ORDER_ENTITY_ID",
  "includeExplanations": true,
  "depth": 2
}

Then call fabric_graph on the sales-order ID. You should see a harmonic edge to the quotation. When includeExplanations is true, explanation.sharedProperties can include values such as customerName or the quote identifier — connections Neural Fabric derived from the properties you supplied, not a relation you wrote.