offdata ai — agentic AI for data modelers and data engineers

Microsoft Fabric

AI Microsoft Fabric schema generator

Describe your domain in plain English. Ship a Fabric Lakehouse or Warehouse — Delta tables in OneLake or T-SQL warehouse tables, plus a complete dbt project.

ddl/fabric.sqlfabric sql
-- Fabric Warehouse (T-SQL)
CREATE TABLE analytics.dim_customer (
  customer_sk   VARCHAR(64)  NOT NULL,
  customer_id   VARCHAR(64)  NOT NULL,
  name          VARCHAR(256),
  plan_id       VARCHAR(64),
  valid_from    DATE         NOT NULL,
  valid_to      DATE
);

-- Fabric Lakehouse (Spark SQL on Delta in OneLake)
CREATE TABLE workspace.lakehouse.fct_usage (
  usage_id      STRING       NOT NULL,
  customer_sk   STRING       NOT NULL,
  plan_id       STRING       NOT NULL,
  event_ts      TIMESTAMP    NOT NULL,
  event_date    DATE         GENERATED ALWAYS AS (CAST(event_ts AS DATE)),
  api_calls     BIGINT       NOT NULL
)
USING DELTA
PARTITIONED BY (event_date);

OffDataAI is purpose-built for Microsoft Fabric. Our generator emits the right DDL for both Fabric modes: Spark SQL on Delta tables when you target a Lakehouse, T-SQL when you target a Warehouse. OneLake three-part names, partition columns generated from event timestamps, and Fabric-friendly type mappings are handled automatically. From the same conversation, you also get a complete dbt project configured for the right Fabric adapter.

What OffDataAI generates

  • Lakehouse + Warehouse modes

    Delta-backed tables for Lakehouse (Spark SQL) and columnstore-friendly DDL for Warehouse (T-SQL). Pick the right mode per project.

  • Kimball, Data Vault 2.0, or 3NF

    Three paradigms supported on Fabric. The generator emits the right syntax for each Fabric mode.

  • Full dbt project

    dbt_project.yml configured for dbt-fabric or dbt-fabricspark, sources, staging, marts, schema tests, and seeds.

  • OneLake-aware identifiers

    Three-part names workspace.lakehouse.table or workspace.warehouse.schema.table — configurable per project in the IR.

Frequently asked questions

Can OffDataAI generate Microsoft Fabric DDL from plain English?
Yes. Describe your business domain in plain English. OffDataAI's interview agent gathers grain, cardinality, and lakehouse vs. warehouse preferences, synthesizes a typed Intermediate Representation, and emits Microsoft Fabric-native DDL — either Lakehouse tables (Delta) or Warehouse tables (T-SQL), with OneLake-aware identifiers.
Does OffDataAI support both Lakehouse and Warehouse modes?
Yes. You can target Fabric Lakehouse (Delta tables in OneLake) or Fabric Warehouse (T-SQL warehouse). The generator emits the right DDL syntax for each — Spark SQL for Lakehouse, T-SQL for Warehouse.
Does it support OneLake three-part names?
Yes. All identifiers use the workspace.lakehouse.table or workspace.warehouse.schema.table format. Catalog and schema names are configurable in the IR.
Does it support Kimball, Data Vault, and 3NF on Fabric?
Yes. All three paradigms are first-class on Fabric. The generator emits Delta-native features in Lakehouse mode and columnstore-friendly DDL in Warehouse mode.
Does it generate a dbt project for Fabric?
Yes. OffDataAI emits a complete dbt project for Fabric — dbt_project.yml configured for dbt-fabric or dbt-fabricspark, sources, staging, marts, schema tests, and seed CSVs.

Your data warehouse is one conversation away.

Describe your domain, or open one of 150+ production-grade templates. ERDs, DDL, and a complete dbt project — generated in under a minute.