Quickstart

Create your first conceptual model, reverse engineer a source, generate a target model, and export DDL.

This guide walks through a first session: create a project, start a free trial, sketch a conceptual model with AI, reverse engineer a source schema, generate a target model, refine it with Copilot, then forward engineer DDL.

Steps

1. Create a project

   Sign in to Snap Data Studio. On the home page, click Create Blank Project.

   Snap Data Studio opens an Untitled Project. You can rename it later from the project menu.

2. Start a free trial

   Agent Copilot and Generate Model need a plan that includes those features.

   1. Open Plans & Billing from the sidebar.
   2. Choose a plan and start the free trial from the pricing table.
   3. Return to your project when billing is active.

   If a gated action asks you to upgrade, use Upgrade to open the same billing page.

3. Use AI to create a conceptual model

   Open the Conceptual Model (use Concepts in the sidebar if you need the Concepts drawer).

   1. Open AI Copilot from the header.
   2. Switch to Agent mode.
   3. Paste the prompt below and send it.
   4. When you see Staged changes ready, review the highlighted concepts and relationships on the canvas.
   5. Click Keep to save the model (or Revert to discard).

   Example prompt:

   Model how we sell parts through customer orders and source those parts from suppliers.
   
   Customers and suppliers each belong to a nation, and every nation sits in a geographic region.
   
   Customers place orders. Each order has line items for specific parts. A part can be supplied by many suppliers, and each supplier can supply many parts — track that part–supplier availability as its own concept.
   

   Prefer business language and singular concept names. You can refine the model later with more Agent prompts or by editing concepts and relationships on the canvas.

   For field-level detail, see Create a conceptual model.

4. Reverse engineer a model

   Import the TPC-H schema as a source ERD.

   1. Open Diagrams in the project sidebar.
   2. Click Import diagram → Import DDL.
   3. Leave the dialect as Snowflake (or choose the dialect that matches the script).
   4. Paste the DDL below into the editor.
   5. Click Generate Diagram.

   TPC-H DDL:

   CREATE OR ALTER TABLE SAMPLE_DATA.TPCH_SF1.PART (
     P_PARTKEY DECIMAL(38, 0) PRIMARY KEY NOT NULL,
     P_NAME VARCHAR(55),
     P_MFGR VARCHAR(25),
     P_BRAND VARCHAR(10),
     P_TYPE VARCHAR(25),
     P_SIZE DECIMAL(38, 0),
     P_CONTAINER VARCHAR(10),
     P_RETAILPRICE DECIMAL(12, 2),
     P_COMMENT VARCHAR(23)
   );
   
   CREATE OR ALTER TABLE SAMPLE_DATA.TPCH_SF1.REGION (
     R_REGIONKEY DECIMAL(38, 0) PRIMARY KEY NOT NULL,
     R_NAME VARCHAR(25),
     R_COMMENT VARCHAR(152)
   );
   
   CREATE OR ALTER TABLE SAMPLE_DATA.TPCH_SF1.NATION (
     N_NATIONKEY DECIMAL(38, 0) PRIMARY KEY NOT NULL,
     N_NAME VARCHAR(25),
     N_REGIONKEY DECIMAL(38, 0),
     N_COMMENT VARCHAR(152),
     FOREIGN KEY (N_REGIONKEY) REFERENCES SAMPLE_DATA.TPCH_SF1.REGION (
       R_REGIONKEY
     )
   );
   
   CREATE OR ALTER TABLE SAMPLE_DATA.TPCH_SF1.CUSTOMER (
     C_CUSTKEY DECIMAL(38, 0) PRIMARY KEY NOT NULL,
     C_NAME VARCHAR(25),
     C_ADDRESS VARCHAR(40),
     C_NATIONKEY DECIMAL(38, 0),
     C_PHONE VARCHAR(15),
     C_ACCTBAL DECIMAL(12, 2),
     C_MKTSEGMENT VARCHAR(10),
     C_COMMENT VARCHAR(117),
     FOREIGN KEY (C_NATIONKEY) REFERENCES SAMPLE_DATA.TPCH_SF1.NATION (
       N_NATIONKEY
     )
   );
   
   CREATE OR ALTER TABLE SAMPLE_DATA.TPCH_SF1.ORDERS (
     O_ORDERKEY DECIMAL(38, 0) PRIMARY KEY NOT NULL,
     O_CUSTKEY DECIMAL(38, 0),
     O_ORDERSTATUS VARCHAR(1),
     O_TOTALPRICE DECIMAL(12, 2),
     O_ORDERDATE DATE,
     O_ORDERPRIORITY VARCHAR(15),
     O_CLERK VARCHAR(15),
     O_SHIPPRIORITY DECIMAL(38, 0),
     O_COMMENT VARCHAR(79),
     FOREIGN KEY (O_CUSTKEY) REFERENCES SAMPLE_DATA.TPCH_SF1.CUSTOMER (
       C_CUSTKEY
     )
   );
   
   CREATE OR ALTER TABLE SAMPLE_DATA.TPCH_SF1.SUPPLIER (
     S_SUPPKEY DECIMAL(38, 0) PRIMARY KEY NOT NULL,
     S_NAME VARCHAR(25),
     S_ADDRESS VARCHAR(40),
     S_NATIONKEY DECIMAL(38, 0),
     S_PHONE VARCHAR(15),
     S_ACCTBAL DECIMAL(12, 2),
     S_COMMENT VARCHAR(101),
     FOREIGN KEY (S_NATIONKEY) REFERENCES SAMPLE_DATA.TPCH_SF1.NATION (
       N_NATIONKEY
     )
   );
   
   CREATE OR ALTER TABLE SAMPLE_DATA.TPCH_SF1.PARTSUPP (
     PS_PARTKEY DECIMAL(38, 0) NOT NULL,
     PS_SUPPKEY DECIMAL(38, 0) NOT NULL,
     PS_AVAILQTY DECIMAL(38, 0),
     PS_SUPPLYCOST DECIMAL(12, 2),
     PS_COMMENT VARCHAR(199),
     PRIMARY KEY (PS_PARTKEY, PS_SUPPKEY),
     FOREIGN KEY (PS_PARTKEY) REFERENCES SAMPLE_DATA.TPCH_SF1.PART (
       P_PARTKEY
     ),
     FOREIGN KEY (PS_SUPPKEY) REFERENCES SAMPLE_DATA.TPCH_SF1.SUPPLIER (
       S_SUPPKEY
     )
   );
   
   CREATE OR ALTER TABLE SAMPLE_DATA.TPCH_SF1.LINEITEM (
     L_LINENUMBER DECIMAL(38, 0) UNIQUE NOT NULL,
     L_LINENUMBE DECIMAL(38, 0) NOT NULL,
     L_SUPPKEY DECIMAL(38, 0),
     L_PARTKEY DECIMAL(38, 0),
     L_QUANTITY DECIMAL(12, 2),
     L_EXTENDEDPRICE DECIMAL(12, 2),
     L_DISCOUNT DECIMAL(12, 2),
     L_TAX DECIMAL(12, 2),
     L_RETURNFLAG VARCHAR(1),
     L_LINESTATUS VARCHAR(1),
     L_SHIPDATE DATE,
     L_COMMITDATE DATE,
     L_RECEIPTDATE DATE,
     L_SHIPINSTRUCT VARCHAR(25),
     L_SHIPMODE VARCHAR(10),
     L_COMMENT VARCHAR(44),
     PRIMARY KEY (L_LINENUMBER, L_LINENUMBE),
     FOREIGN KEY (L_LINENUMBE) REFERENCES SAMPLE_DATA.TPCH_SF1.ORDERS (
       O_ORDERKEY
     ),
     FOREIGN KEY (L_PARTKEY, L_SUPPKEY) REFERENCES SAMPLE_DATA.TPCH_SF1.PARTSUPP (
       PS_PARTKEY,
       PS_SUPPKEY
     )
   );
   

   Snap Data Studio creates a new diagram from the script. It does not overwrite your conceptual model.

   For other import options (database or DBML), see Create a logical & physical model.

5. Generate a logical and physical model

   Combine your conceptual model with the reverse-engineered source to produce a target ERD.

   1. With the Conceptual Model or source diagram tab active, click Generate Model in the header.
   2. Choose a methodology — for analytics, pick Kimball - Dimensional Model.
   3. On Concepts, select the concepts to include (use Select All for this walkthrough).
   4. On Source Model, select the diagram you imported from DDL.
   5. Optionally add guidance on the Guidance tab.
   6. Click Generate Model and wait for the run to finish.

   Generation always creates a new diagram (for example the source name with a - Dimensional suffix), plus AI-Generated Documentation and often Column Mappings. It does not overwrite the source diagram.

   Open the new diagram and use the Logical / Physical toggle to switch views. If physical mode is incomplete, set Warehouse, Database or Catalog, and Schema from physical location settings.

   For methodology options and how concepts enrich generation, see Create a logical & physical model.

6. Refine the model with Copilot

   Use Agent mode on the generated diagram to make a real modeling change — promote shipping method into its own dimension.

   1. Open the generated dimensional diagram.
   2. Open AI Copilot and switch to Agent mode.
   3. Paste the prompt below and send it.
   4. Review the staged table and relationship changes on the canvas.
   5. Click Keep to apply them (or Revert to discard).

   Example prompt:

   Add a Ship Mode dimension for how line items are shipped. Link it to the line item fact, give it a surrogate key, and include the ship mode code and a display name. If ship mode still exists only as an attribute on the fact, remove that attribute after the dimension is in place.
   

   Staged changes are not saved until you Keep. Resolve them before starting another Agent edit.

   For more Copilot patterns on diagrams, see Create a logical & physical model.

7. Export an artifact

   Forward engineer the generated diagram into warehouse DDL.

   1. Open Diagrams in the project sidebar.
   2. Hover the generated diagram and open its ⋯ menu.
   3. Click Forward engineer.
   4. Confirm the database type (Snowflake or Databricks).
   5. Click Generate DDL.

   A .sql artifact opens under the diagram. Edit and autosave as needed — Snap Data Studio does not run the script against your warehouse.

   You can also add documentation from the same ⋯ menu, or review the docs generation already created.

Next steps

• Connect to a data warehouse
• Create a conceptual model
• Create a logical & physical model
• Create an artifact