A data model is the agreement between the people who own the data and the people who build on it. When that agreement lives only in someone’s head or in a stale wiki diagram, every new pipeline, report or AI feature has to rediscover what a “customer” or an “active account” actually means. Data modeling tools turn that knowledge into conceptual, logical and physical models that can be versioned, reviewed and pushed to a real database as DDL.
The category has changed a lot in the last few years. The long-running desktop suites have added Git integration, cloud repositories and AI assistants. Browser-based modelers built for Snowflake, Databricks and BigQuery now compete for the same teams, and polyglot tools handle JSON, Avro and NoSQL schemas that classic ER diagram tools never touched. If you are moving to a cloud warehouse at the same time, our guide on the best way to migrate from a data warehouse to a lakehouse is worth reading alongside this list.
We ranked these ten tools on modeling depth (logical and physical models, standards, glossaries), database coverage, collaboration and version control, reverse engineering quality and how transparent each vendor is about pricing.
Data Modeling Tools At A Glance
Tool
Best For
Model Levels
Deployment
Pricing Model
erwin Data Modeler
Large enterprises tying models to data governance
Conceptual, logical, physical
Desktop with on-prem or cloud Mart
Quote-based
ER/Studio
Enterprise architecture teams that need a shared repository and glossary
Logical, physical, enterprise metadata
Desktop plus repository and Team Server
Standard and Pro bought online, Enterprise by quote, free trial
SqlDBM
Cloud data teams on Snowflake, Databricks and dbt
Conceptual, logical, physical, semantic
Browser-based SaaS
Per seat, quote-based
Hackolade Studio
NoSQL, JSON, Avro and API schema design
Conceptual, logical, physical (polyglot)
Desktop or browser
Subscription per seat, free browser edition
Navicat Data Modeler
Cross-platform desktop modeling with Data Vault support
Conceptual, logical, physical
Desktop (Windows, macOS, Linux)
Perpetual license, free Lite edition
DbSchema
Git-based schema design across 100+ databases
Logical, physical
Desktop
Free Community Edition, paid Pro
Redgate Data Modeler
Small teams that want a low-cost browser modeler
Logical (Pro), physical
Browser-based SaaS
Published annual per-user pricing, 14-day trial
Toad Data Modeler
Windows shops modeling Oracle, SQL Server and Db2
Logical, physical
Windows desktop
Quote-based, free trial
Oracle SQL Developer Data Modeler
Oracle teams that need a free modeler
Logical, relational, dimensional
Desktop (Windows, Linux, macOS)
Free
Lucidchart
Quick collaborative ER diagrams for mixed audiences
Diagram level
Browser-based SaaS
Free plan, paid tiers
Best Data Modeling Tools In 2027
1. erwin Data Modeler — Models Linked To Governed Metadata Through erwin Data Intelligence
erwin Data Modeler, now part of Quest Software, is one of the oldest names in the category and still one of the most complete. It supports reverse engineering, forward engineering, normalization and denormalization, bulk editing and full model compare. Database coverage is wide: Oracle, SQL Server, MySQL, PostgreSQL, Db2, Teradata, MariaDB, Snowflake, Databricks, Google BigQuery, Apache Hive, MongoDB, Cassandra, Neo4j, plus Parquet and JSON formats.
What sets erwin apart is how far the model travels after it is drawn. Integration with erwin Data Intelligence and the Quest data management platform links entities and attributes to a business glossary and governed metadata, so a column definition in the model can show up in the catalog that analysts search. Recent releases added Git version control, GenAI-assisted model and schema generation and erwin ER360, a web portal for sharing models with people who do not run the desktop tool.
Editions include Standard and a Workgroup edition with a central repository (the Mart), which can now run in the cloud. Pricing is not published. The trade-off is cost and complexity: erwin rewards organizations with formal data governance programs, but a five-person analytics team may find it heavier than it needs.
Best for: enterprises that want data models to feed a governed metadata and glossary program.
2. ER/Studio — Shared Model Repository With Business Glossary And Purview And Collibra Integration
ER/Studio, from Idera, covers logical and physical modeling with automated logical-to-physical transformation, forward engineering with DDL generation, reverse engineering of existing databases, compare and merge, macros and Git integration. Supported platforms include SQL Server, Oracle, PostgreSQL, Snowflake, Databricks, Microsoft Fabric, Google BigQuery, Amazon Redshift and MongoDB.
The product is sold in clear tiers. Data Architect handles individual modeling with a local data dictionary. Data Architect Pro adds a central repository with version history, check-in and check-out and an enterprise data dictionary for teams. Enterprise bundles those with Team Server Core and adds a web portal, metadata import and export, a business glossary and data catalog, and integrations with Microsoft Purview and Collibra. Business rules and naming standards can be enforced across models, which matters when dozens of modelers share one repository. An AI assistant, ERbert, helps with modeling tasks.
A free trial is available, the Standard and Pro editions can be bought online, and Enterprise is quote-based. Like erwin, ER/Studio is a desktop-centered enterprise suite, so teams that live entirely in the browser may prefer a SaaS modeler. Its strength is the combination of a repository, naming standards and governance integrations at each tier.
Best for: data architecture teams that need repository-based collaboration and a business glossary connected to Purview or Collibra.
3. SqlDBM — Browser-Based Modeling With dbt YAML Export For Snowflake And Databricks
SqlDBM runs entirely in the browser with nothing to install, and it was built for the cloud data stack. Supported platforms include Snowflake, Databricks, Google BigQuery, Azure Synapse, Amazon Redshift and Microsoft Fabric, along with SQL Server, Oracle, MySQL, PostgreSQL and Teradata. It handles conceptual, logical and physical models, reverse and forward engineering, and model-to-database synchronization.
The dbt integration is the reason many analytics engineering teams pick it. Models can be exported as dbt YAML and pushed, along with generated DDL, straight to GitHub. Column-level lineage, impact analysis, naming standards and change review workflows round out the governance side, and a semantic layer lets teams define governed metrics for BI and AI consumers. An AI Copilot and MCP access for AI agents are also available, with AI usage priced separately from seats.
SqlDBM prices per seat with no feature tiers, but the price is quote-based. It is less suited to teams that model on-premises transactional systems or need an offline tool. For Snowflake-heavy shops, it pairs well with the partners in our list of Snowflake consulting partners.
Best for: analytics engineering teams on Snowflake or Databricks who work in dbt and Git.
4. Hackolade Studio — Polyglot Models That Generate Schemas For 45-Plus Targets Including NoSQL And Avro
Hackolade Studio is a polyglot modeler. You define a technology-agnostic model once and derive physical schemas for more than 45 targets: relational databases such as Oracle, PostgreSQL and SQL Server; cloud analytics platforms such as Snowflake, Databricks, BigQuery and Redshift; NoSQL stores such as MongoDB, Cassandra, DynamoDB and Neo4j; and schema formats including JSON Schema, Avro, Parquet, Protobuf, OpenAPI and GraphQL.
That makes it the strongest choice when the same business entity has to exist as a Postgres table, a Kafka Avro message and an API payload. It reverse-engineers live databases, DDL, JSON, Excel and even PowerDesigner and erwin models, and it forward-engineers DDL, validators and documentation. Models stay on your machine or in your Git repositories rather than on Hackolade servers.
Hackolade runs as a desktop app or in the browser, and a free browser edition is available. Paid editions are subscription-based per seat, with floating seats on request, and pricing is published on the vendor site. Teams that only model relational warehouses may not need the polyglot depth.
Best for: platform and API teams doing NoSQL data modeling and event schema design.
5. Navicat Data Modeler — Data Vault 2.0 And Dimensional Modeling On Windows, macOS And Linux
Navicat Data Modeler is a desktop tool that runs on Windows, macOS and Linux, which is less common than it should be in this category. It handles conceptual, logical and physical models in one workspace, converts between levels, and supports relational, dimensional and Data Vault 2.0 methods with Crow’s Foot, IDEF1X and UML notations.
Supported databases include MySQL, PostgreSQL, MariaDB, SQL Server, Oracle, SQLite, MongoDB and Snowflake, plus cloud variants on Amazon RDS, Azure and Google Cloud SQL. Reverse engineering works from a live connection or from pasted or imported DDL, and forward engineering exports all or part of a model as SQL. Newer features include a data dictionary with PDF export, naming standards, a shared glossary, model comparison and AI-assisted table design.
Navicat sells a perpetual license and offers a free Lite edition. It suits individuals and small teams well, though it lacks the enterprise metadata and governance integrations of erwin or ER/Studio.
Best for: Mac and Linux users, and warehouse modelers working in Data Vault 2.0.
6. DbSchema — Git-Friendly Text Model Files And Offline Design For 100-Plus Databases
DbSchema turns more than 100 SQL and NoSQL databases into ER diagrams and interactive documentation. Main engines such as MySQL, PostgreSQL, Oracle, SQL Server, MariaDB, SQLite, Snowflake, MongoDB, Redshift and BigQuery have native connectors, and others connect through JDBC.
Its design model is stored as plain-text files, so teams can branch, diff and review schema changes in pull requests like any other code. Schemas can be designed offline and later compared with a live database, with the tool showing the SQL for each change before it runs.
A free Community Edition covers non-commercial use, and a 15-day Pro trial is included. It is closer to a developer’s database design software than a governance suite, so it does not offer enterprise glossaries or metadata catalogs.
Best for: developers who want schema design and review inside a Git workflow.
7. Redgate Data Modeler — Former Vertabelo With Published Per-User Pricing And Real-Time Collaboration
Redgate acquired the Vertabelo data modeler in September 2025 and relaunched it as Redgate Data Modeler. It is a browser-based tool with real-time collaboration, versioning and change history, reverse engineering, live validation and automatic generation of create, drop and partial change scripts. Supported databases are SQL Server, Oracle, MySQL, PostgreSQL, SQLite, Db2, Snowflake, Amazon Redshift, HyperSQL and Google BigQuery.
Pricing is refreshingly clear. The Standard plan covers physical design with up to 20 models and 100 tables per model, Pro adds logical design, unlimited models and API access, and both are priced per user per year on Redgate’s site. It is also bundled into Redgate’s SQL Toolbelt Essentials, and a 14-day trial is available. The model limits on the entry plan and the shorter database list keep it below the leaders here.
Best for: small and mid-size teams that want a simple ER diagram tool in the browser with predictable cost.
8. Toad Data Modeler — Cross-Database Structure Migration And Model Sync For Windows Shops
Toad Data Modeler is Quest’s lighter-weight modeler, separate from erwin. It builds logical and physical ER models, reverse and forward engineers databases, generates DDL, compares and synchronizes models, and migrates structures from one platform to another. Coverage includes Oracle, SQL Server, Db2 (LUW and z/OS), PostgreSQL, MySQL, Teradata, SQLite, SAP ASE, SAP IQ, Greenplum and others, with a Universal Model for anything not listed.
Version 8.0, released in March 2025, added support for Oracle 21c, SQL Server 2022 and Windows 11. It is a Windows desktop application, and the release notes list support only up to PostgreSQL 12 and MySQL 8.0, so check version coverage if you run newer engines. A free trial is available and pricing is by quote.
Best for: Windows-based database teams modeling Oracle, SQL Server and Db2.
Oracle SQL Developer Data Modeler is free, which alone earns it a place on any list of data modeling tools. It supports entity relationship diagrams, relational, data type and multi-dimensional models, forward and reverse engineering with DDL generation, design validation against built-in rules, custom design rules and transformations, Subversion integration and built-in reports.
It runs on Windows, Linux and macOS (including Apple silicon). The 24.3.1 release notes list import from Oracle Database, SQL Server, Db2 LUW and Db2 for z/OS, with script export for Oracle, SQL Server and Db2. Non-Oracle drivers have to be added manually.
The trade-off is focus. It is excellent for Oracle work and acceptable for SQL Server and Db2, but it does not target cloud warehouses or NoSQL, and collaboration relies on Subversion rather than Git.
Best for: Oracle Database teams that need a capable modeler at no cost.
10. Lucidchart — Real-Time Collaborative ER Diagrams With SQL Export For Four Major Databases
Lucidchart is a general diagramming platform rather than a dedicated modeler, but its ER diagram features cover a lot of early-stage design work. You can import database tables and schemas (and Salesforce schemas) to build a diagram, let Lucidchart draw relationship lines from the structure, and export the finished ERD as SQL for MySQL, PostgreSQL, SQL Server or Oracle.
Its advantage is the audience. Product managers, analysts and engineers can edit the same diagram in real time without installing anything, which makes it useful for workshops and conceptual models that non-specialists need to review. A free plan is available, with paid Individual, Team and Enterprise tiers.
Lucidchart is not built around logical-to-physical transformation, naming standards or model compare, so it complements rather than replaces a dedicated tool once a design moves toward production.
Best for: collaborative conceptual diagrams shared with business stakeholders.
How To Choose A Data Modeling Tool
Which Platforms Will Your Models Target?
List the databases, warehouses and schema formats you need to generate today and in two years. Classic enterprise suites cover the widest relational spread, SqlDBM is tuned for cloud warehouses, and Hackolade leads for NoSQL data modeling and message schemas. A tool that cannot emit DDL for your main platform becomes a drawing tool.
Do You Need Logical And Physical Models, Or Just Diagrams?
Teams that maintain separate logical and physical models, with naming standards that translate business names into column names, need a real modeler such as erwin, ER/Studio, SqlDBM or Navicat. If the goal is a shared picture for a design review, a lighter ER diagram tool may be enough.
How Will Models Be Versioned And Reviewed?
Some tools use a central repository with check-in and check-out, others store models as files in Git, and SaaS tools keep version history in the cloud. Git-based workflows suit teams practicing CI/CD for databases, while repositories suit large modeling groups with formal approvals.
How Good Is Reverse Engineering On Your Existing Databases?
Most projects start from an existing schema, not a blank canvas. Trial each shortlisted tool against your largest production database and check how it handles views, constraints, comments and platform-specific types.
Does The Model Need To Feed Governance Or BI?
If your data catalog, glossary or semantic layer should inherit definitions from the model, prioritize tools with integrations for Purview, Collibra, Atlan or dbt. If your modeling work supports reporting projects, our overview of business intelligence companies and service providers can help you find implementation partners.
Conclusion
The best data modeling tools in 2027 split along two lines: where the model lives (desktop repository, Git or the browser) and how far it travels after it is drawn. erwin and ER/Studio remain the leaders for enterprises that tie logical and physical models to glossaries and governance programs. SqlDBM is the natural fit for dbt-centered cloud teams, and Hackolade is the specialist for NoSQL and event schemas.
For smaller budgets, Navicat, DbSchema and Redgate Data Modeler offer solid modeling with clearer pricing, while Oracle’s free modeler and Lucidchart cover narrower needs well. Run reverse engineering on a real schema during the trial, and choose the tool your modelers and engineers will both keep using.
