offdata ai — agentic AI for data modelers and data engineers

An AI data modeling platform

Describe your domain.
Ship a warehouse.

OffDataAI turns a plain-language description of your business into a validated, production-ready data model — ERDs, DDL, a full dbt project, and seed data — for every major cloud warehouse. No drag-and-drop. No blank canvas. No SQL required.

Get startedBook a demoSee how it works

Ships native models for

  • Snowflake
  • BigQuery
  • Databricks
  • Redshift
  • Postgres
  • Synapse
  • Fabric
  • dbt
  • Snowflake
  • BigQuery
  • Databricks
  • Redshift
  • Postgres
  • Synapse
  • Fabric
  • dbt
new-project.offdataai.cominterview · sessionir.jsonerd · star schemaddl + dbt + seed
● live
Describe your domain

We're a B2B SaaS that meters API usage per workspace. Customers buy subscriptions, each with multiple seats. Bills are generated monthly from usage rollups and plan tier.

243 charsStart interview →

interview agent

Got it — multi-tenant SaaS with subscriptions and metered usage. A couple of clarifying questions:

interview agent

What's the grain of a usage event — per API call, per session, or per feature invocation?
Per API call. We meter every request.

interview agent

Should the customer dimension be SCD2 — keep history when plan changes?
Yes, SCD2 on plan, SCD1 on contact info.
ir.jsonvalidated
{
  "paradigm": "kimball",
  "entities": [
    { "name": "dim_customer", "scd": "type_2" },
    { "name": "dim_plan",     "type": "dimension" },
    { "name": "fct_usage",    "grain": "api_call",
      "fk": ["customer_sk", "plan_id"] },
    { "name": "fct_invoice",  "grain": "invoice_line",
      "derived_from": ["fct_usage", "dim_plan"] }
  ],
  "relationships": [
    { "from": "fct_usage",   "to": "dim_customer" },
    { "from": "fct_invoice", "to": "fct_usage"   }
  ]
}
erd · star schema

ddl/snowflake.sql

create table fct_usage (
  usage_sk     varchar,
  customer_sk  varchar,
  plan_id      varchar,
  event_ts     timestamp,
  api_calls    number
)
cluster by (event_ts);

dbt project

▾models/
▾staging/
·stg_customer.sql
·stg_usage.sql
▾marts/
·dim_customer.sql
·fct_usage.sql
·fct_invoice.sql
▾seeds/
·plans.csv
01Describe your domain02AI interview03IR synthesis04ERD generation05Artifacts emitted

Built by data architects, for data teams

  • Analytics engineers

    Ship a warehouse without writing every table

  • Data architects

    Evaluate paradigms against the same domain

  • Founding teams

    Stand up a data layer without a data team

  • Consultancies

    Deliver across many client domains

What you ship

One IR. Every artifact your warehouse needs.

From a single validated Intermediate Representation, OffDataAI emits the full delivery package — ERD, platform-native DDL, a scaffolded dbt project, and realistic seed data — wired together and ready for review.

ERDStar, snowflake, or vault

Mermaid diagram

DDL7 warehouses · clustering & partitions
create table fct_orders (
  order_id     varchar,
  customer_sk  varchar,
  product_sk   varchar,
  order_ts     timestamp,
  amount       number
)
cluster by (order_ts);

Platform-native SQL

dbtstaging · marts · tests
▾models/
▾staging/
·stg_customer.sql
·stg_orders.sql
▾marts/
·dim_customer.sql
·dim_product.sql
·fct_orders.sql
·_schema.yml

Scaffolded project

SeedPer table, honouring relationships
customer_id,name,tier
C-1001,Acme Corp,pro
C-1002,Globex,team
C-1003,Initech,starter
C-1004,Hooli,enterprise
C-1005,Pied Piper,team
C-1006,Stark Inds,enterprise

Realistic CSVs

What it is

From conversation to production schema

Four pillars that make OffDataAI different from every other data-modeling tool.

  • 01

    Conversational modeling

    Describe your business in plain English. An interview agent asks targeted follow-up questions — grain, cardinality, SCDs — until the model is unambiguous.

  • 02

    Every paradigm, every platform

    Kimball star schema, Data Vault 2.0, or 3NF — targeting Snowflake, Databricks, Postgres, BigQuery, Redshift, Synapse, or Fabric. One conversation, any destination.

  • 03

    Complete artifact generation

    From a single IR, generate Mermaid ERDs, platform-specific DDL, a fully scaffolded dbt project with staging and marts, and realistic seed data — wired together.

  • 04

    An open IR contract

    Everything flows through a validated JSON Intermediate Representation. Inspect it, patch it, version it. Your data model is never a black box.

How it works

Try the flow — click through a sample build

From a blank prompt to deployable artifacts. Walk through each step to see exactly what happens.

Step 1 of 4

Describe your domain

Tell us about your business in plain language — no schema knowledge required.

new-project.offdataai.com
We're a B2B SaaS platform that tracks subscriptions, usage events, and billing across multi-tenant customers. Customers have plans, plans meter on usage, and we invoice monthly.
177 chars

Try a different example

The product

Everything you need to model, generate, and ship

Interactive Modeling

Chat-driven schema design

No drag-and-drop. No manual table creation. Just describe what your business does, answer a few questions, and watch the schema take shape in real time.

  • Natural language input
  • Smart follow-up questions
  • Real-time IR preview

interview · session

We run a B2B SaaS — customers buy subscriptions, each with multiple seats.
Got it. Do seats roll up to a single billing account, or can a customer have multiple?
One billing account per customer, but multiple workspaces underneath.
drafting schema…

ERD Visualization

Every relationship, one diagram

Automatically generated Mermaid ERDs that update as your model evolves. See dimensions, facts, hubs, links, and satellites at a glance.

  • Auto-generated Mermaid diagrams
  • Exportable SVG/PNG
  • Paradigm-aware layouts

erd · star schema

dim_customer
customer_id
name
tier
dim_plan
plan_id
name
price
fct_subscriptions
customer_id FK
plan_id FK
date_id FK
mrr_amount
dim_date
date_id
day, month, qtr

DDL Generation

Platform-native SQL, instantly

Generate CREATE TABLE statements tuned for your target — Snowflake clustering keys, BigQuery partitioning, Postgres constraints. Copy, run, done.

  • 7 target platforms
  • Platform-specific optimizations
  • One-click copy
SnowflakeBigQueryPostgres
fct_subscriptions.sql
CREATE TABLE analytics.fct_subscriptions (
  subscription_id   VARCHAR(36)  NOT NULL,
  customer_id       VARCHAR(36)  NOT NULL,
  plan_id           VARCHAR(36)  NOT NULL,
  date_id           DATE         NOT NULL,
  mrr_amount        NUMBER(18,2),
  PRIMARY KEY (subscription_id)
)
CLUSTER BY (date_id, customer_id);
Copy DDL

dbt Projects

Staging, marts, and models — scaffolded

Get a fully structured dbt project with source definitions, staging models, and mart layers. Ready for dbt build from day one.

  • Source YAML definitions
  • Staging + mart layers
  • Realistic seed data

project tree

dbt build ready
▾my_project/
·dbt_project.yml
▾models/
▾staging/
·stg_customers.sqlSQL
·stg_subscriptions.sqlSQL
·_sources.yml
▾marts/
·dim_customer.sqlSQL
·dim_plan.sqlSQL
·fct_subscriptions.sqlSQL
▾seeds/
·customers.csv

Templates · 150+ and counting

Don't start from a blank canvas. Start from a warehouse.

Every template is a complete, validated data model — entities, attributes, keys, PII tags, SCDs, and relationships already wired. Pick your industry, open a fully-populated ERD, and generate DDL plus dbt in one click.

  • Kimball73

    Star schemas with conformed dimensions and fact grains for analytics.

  • Data Vault 2.039

    Hubs, links, and satellites with hash keys for enterprise warehouses.

  • 3NF40

    Normalized operational models that mirror source-system schemas.

Spanning every industry

Banking & FinanceHealthcareInsuranceSaaSRetail & E-commerceMarketingSAPSalesforceTelecomManufacturingLogisticsHR & WorkforcePublic SectorEducationHospitalityGamingAviationEnergy & Utilities+ many more
Browse all templatesUse a template

The platform

Everything a data team needs to ship — and govern — a warehouse

Generation is just the start. OffDataAI takes the model all the way to production: into your repo, your warehouse, your team, and your compliance review.

  • Import an existing warehouse

    Connect Postgres (more coming), introspect the live schema, and start from real tables — then refine and regenerate.

  • One-click push to Git

    Commit the generated DDL, dbt project, seed SQL, and docs straight to GitHub — and open a PR — so it lands in your real workflow.

  • Governance and compliance

    Classify columns, flag PII, set masking per attribute. Generate masking-policy DDL and a compliance report — built for regulated industries.

  • Schema migrations

    Diff any two model versions and get the exact ALTER / CREATE migration SQL to evolve a live warehouse — no hand-written DDL, no drift.

  • Team collaboration

    Invite teammates as viewers or editors, share a read-only link with stakeholders, and discuss the model with comments tied to entities.

  • Public API and keys

    Mint API keys and create projects, fetch the IR, or pull DDL programmatically — wire OffDataAI straight into your CI and internal tooling.

  • 150+ ready templates

    Start from a complete domain model — banking, healthcare, insurance, SaaS, retail, SAP, Salesforce — pre-filled, validated, ready to refine.

  • Email and Slack notifications

    Synthesis runs in the background. Get an email or Slack message the moment your model is ready, or if a run needs your attention.

Platforms & paradigms

One model, every destination

Generate platform-native DDL and a complete dbt project for every major cloud data platform — and pick the modeling paradigm that fits the work.

  • Snowflake

    Clustering keys, transient tables

  • Databricks

    Delta Lake, Unity Catalog

  • PostgreSQL

    Constraints, indexes, partitions

  • BigQuery

    Partitioning, clustering

  • Redshift

    Distribution, sort keys

  • Synapse

    Dedicated SQL pools

  • Fabric

    Lakehouse tables

  • Your integration

    Don't see your stack? Tell us.

Supported modeling paradigms

  • Most popular

    Kimball

    Star schema with dimensions and facts

  • Enterprise

    Data Vault 2.0

    Hubs, links, and satellites for warehouses at scale

  • Operational

    3NF

    Normalized relational modeling for source systems

FAQ

Everything you wanted to ask

What is OffDataAI?
OffDataAI is an AI data modeling tool that converts a natural-language description of your business domain into a production-ready data warehouse. It generates ERDs, platform-native DDL, complete dbt projects, and realistic seed data — for Snowflake, BigQuery, Databricks, Redshift, Postgres, Synapse, and Microsoft Fabric.
How is OffDataAI different from ChatGPT or a generic LLM for data modeling?
OffDataAI is a purpose-built pipeline: an interview agent gathers grain, cardinality, and SCD requirements; a synthesis agent compiles your answers into a validated Intermediate Representation (IR); validators check referential integrity and type coercions before any code is generated; and platform-specific generators emit DDL, dbt scaffolds, and seed CSVs. ChatGPT can sketch a schema, but OffDataAI ships warehouse-ready artifacts that are tested and consistent across paradigms (Kimball, Data Vault 2.0, 3NF).
What is the Intermediate Representation (IR)?
The IR is a validated JSON document that captures your entire data model — entities, attributes, relationships, grain, and platform hints. It is the single source of truth that every generator consumes. You can inspect, edit, and version it like any other artifact.
Which LLM models does OffDataAI use?
We use Claude Sonnet 4.5 for synthesis (turning your conversation into a structured IR) and Claude Haiku 4.5 for the interview agent (fast, conversational follow-ups). All LLM calls go through Anthropic's API.
Can I edit the generated schema after it's created?
Yes. The IR is fully editable — you can patch it through the API or the UI. Changes flow through the validators and regenerate downstream artifacts automatically.
Which data warehouses and databases does OffDataAI support?
OffDataAI ships native DDL and dbt projects for Snowflake, Google BigQuery, Databricks (Delta), Amazon Redshift, PostgreSQL, Microsoft Synapse, and Microsoft Fabric. Each generator respects platform-specific syntax, clustering, partitioning, and data type rules.
Does OffDataAI support Kimball, Data Vault, and 3NF modeling?
Yes. OffDataAI supports three modeling paradigms out of the box: Kimball dimensional modeling (star and snowflake schemas), Data Vault 2.0 (hubs, links, and satellites), and Third Normal Form (3NF). You choose the paradigm — the synthesis agent shapes the IR accordingly.
What does a generated dbt project include?
A complete dbt project with source YAML definitions, staging models, mart models, schema tests, and realistic seed CSV data. It is ready for `dbt build` from day one.
Is my data sent to any third-party service?
Your domain descriptions are sent to Anthropic's Claude API for processing. No data is stored on third-party servers beyond what is needed for the API call. All generated artifacts are stored in your own PostgreSQL database and object storage (MinIO or Cloudflare R2).
Can I import my existing data warehouse?
Yes. Connect a database (PostgreSQL today, more warehouses coming) and OffDataAI introspects the live schema — tables, columns, primary keys, and foreign keys — and seeds a new project with an editable IR. You can then refine the model in the interview and regenerate clean DDL, dbt, and docs from what you already have.
Can OffDataAI push the generated code to my Git repository?
Yes. Connect GitHub and push the generated DDL, dbt project, seed SQL, documentation, and the canonical IR to a branch in one click — optionally opening a pull request — so the output is reviewed and version-controlled in your real workflow.
Does OffDataAI handle data governance and PII?
Yes. Every column carries a classification and an optional masking strategy. OffDataAI inventories PII, reports masking coverage and compliance tags (e.g. HIPAA, GDPR), and generates platform-appropriate masking-policy DDL — for example Snowflake masking policies — so regulated teams can enforce governance from day one.
Can I generate migrations when my model changes?
Yes. Every edit creates a new IR version. OffDataAI diffs any two versions and generates the exact ALTER TABLE / CREATE TABLE migration SQL for your target platform, so you can evolve a live warehouse safely instead of hand-writing DDL.
Can my team collaborate on a data model?
Yes. Invite teammates as viewers or editors, share a read-only link with stakeholders who don't have an account, and leave inline comments on the model for review and sign-off.
Is there an API?
Yes. Create an API key and use the public REST API to create projects, fetch the generated IR, and pull platform-native DDL programmatically — ideal for wiring OffDataAI into CI pipelines and internal tooling.
Are there templates to start from?
Yes — 136 of them. The template gallery offers complete, validated domain models across 18+ industries (banking, healthcare, insurance, SaaS, retail, SAP, Salesforce, telecom, and more), spanning all three paradigms: 61 Kimball star schemas, 39 Data Vault 2.0 warehouses, and 36 normalized 3NF models. Selecting one instantiates a project with a fully-populated ERD so you can ship a warehouse for a common domain in one click and refine from there.
How do I know when my model is ready?
Synthesis runs in the background. OffDataAI can notify you by email and/or Slack the moment your model is ready — or if a run needs attention — so you don't have to watch a progress bar.
Is there a free tier?
There's no self-serve free tier — but we're offering OffDataAI completely free to our first customers. If you'd like a free subscription, just connect with the admins of the tool and they'll provide you with a free access code. Reach out at sarora@s2datasystems.in or book a demo and we'll get you set up.

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.

Get startedTalk to usBrowse templates