MetaFabric

The data and knowledge foundation for your applications and AI.

Business data is spread across many systems, and each one describes it differently. That makes it hard to trust, hard to build applications on, and hard for AI to use safely.

MetaFabric connects those systems, brings their data into one shared business model, and serves it to people, applications and AI agents. It is more than an integration layer: applications can use it as their system of record and read and write their data directly. Data keeps its link to the system it came from, and only the people and programs you allow can reach it.

Figure 1. MetaFabric sits between the systems a business runs and everything that uses its data.

Figure 1. MetaFabric sits between the systems a business runs and everything that uses its data.

What it does#

Connects your systems

Salesforce, ServiceNow, Jira, NetSuite, Workday, Supabase and custom sources.

Unifies the data

One business model across systems. An account is an account, wherever it came from.

Powers applications

A governed system of record that apps read and write through APIs.

Enables AI and agents

Answers from your documents with sources, MCP access for AI assistants, and agents that act after approval.

Validates data

Checks loaded data against its source and flags what needs fixing.

Controls access

Separate workspaces, roles, object permissions and record-level rules.

How it works#

There are three ways to use MetaFabric, and a workspace can combine them.

Approach	What happens
Read live	Connect a system and map it to the shared model. Records stay in the source and are read when needed.
Bring in and unify	Copy a system's records in and combine them with other sources. A person approves before it publishes, and checks confirm it is complete.
Build on it	Applications create and manage their own records in MetaFabric through APIs.

In every case the same access rules and audit apply.

Where to start#

You are	Read
A leader evaluating MetaFabric	Capabilities at a glance, then Canonicalization
An architect	Architecture and services, Security, access and tenancy
A developer building an app	Application data foundation, Developer APIs, MCP
An admin setting it up	Connecting source systems, Security, access and tenancy

Key terms#

Term	Meaning
Workspace	An isolated space with its own data, users and settings.
Tenant	The same isolated space, as named in the APIs. Each workspace is one tenant.
Canonical ontology	The shared business vocabulary (`ec:`) all data is described in.
Crosswalk	The map from each system's objects and fields to the canonical ontology.
Mapping	The reviewed process of bringing one system's data into MetaFabric.
Knowledge fabric	The connected picture of a workspace's objects, their links and their sources.
System of record	The system trusted as the source of truth for a field.