Can I use standard Postgres tools and drivers with a lakehouse-backed database?

Summary

• Yes. Databricks Lakebase speaks the standard PostgreSQL wire protocol, so any standard Postgres client, driver, or ORM connects to it without custom adapters.
• Your existing tooling carries over: command-line clients like psql, GUI clients like pgAdmin and DBeaver, and language drivers and ORMs across Python, Java, .NET, JavaScript, and Go.
• Connections use a normal Postgres connection string over TLS. You authenticate with a native Postgres password or a short-lived OAuth token tied to a Databricks identity or service principal.
• Lakebase is built on Neon, the serverless Postgres engine Databricks acquired in 2025, and supports PostgreSQL 16, 17, and 18 plus a wide range of Postgres extensions.
• A few managed-service limits apply: no operating-system or Postgres superuser access, no tablespaces, and native logical replication isn't available yet.

Can I use standard Postgres tools and drivers with a lakehouse-backed database?

Yes. Databricks Lakebase is a Postgres database that exposes the standard PostgreSQL wire protocol, so the client tools, drivers, ORMs, and SQL you already use with Postgres connect to it without change. "Lakehouse-backed" describes how Lakebase integrates with the rest of the Databricks platform, not a different connection surface: to your application and your tooling, it looks and behaves like Postgres. Your team's existing Postgres skills, libraries, and scripts carry over directly.

What "lakehouse-backed" changes, and what it doesn't

Lakebase is a fully managed, serverless Postgres database built on Neon, the serverless Postgres engine Databricks acquired in 2025, and it runs inside the Databricks Data Intelligence Platform under Unity Catalog governance. The lakehouse integration shows up in what surrounds the database: unified governance, analytics on the same data, and syncing with lakehouse tables. It does not change the connection surface. The protocol, SQL dialect, drivers, and connection string stay standard Postgres, which is exactly what lets your current tools work.

What connects out of the box

Tool or driver category	Works with Lakebase
Command-line clients	psql and other native PostgreSQL CLIs.
GUI clients	Standard Postgres GUIs such as pgAdmin and DBeaver.
Language drivers	Standard Postgres drivers, including psycopg for Python, JDBC and ODBC for Java and .NET, and the native Go driver.
ORMs and frameworks	Popular ORMs such as SQLAlchemy, Django, and Prisma connect through their usual Postgres driver.
Extensions	A wide range of Postgres extensions, installed with standard SQL.

Databricks publishes copy-paste connection snippets for these languages and frameworks, and the Lakebase UI has a Connect dialog that generates them for you.

How Databricks approaches Postgres compatibility

Connections use an ordinary Postgres connection string of the form postgresql://role:password@host/database, and Lakebase requires TLS, so sslmode=require is part of the standard setup. You authenticate one of two ways: native Postgres password authentication, which the documentation recommends for tools and long-running processes, or a short-lived OAuth token tied to a Databricks identity or service principal that your application rotates automatically.
On versions and extensions, Lakebase supports PostgreSQL 16, 17, and 18, with 17 as the default, and you choose the version when you create a project. It supports a wide range of Postgres extensions installed with standard SQL. For AI applications, Lakebase offers vector search through an interface that is drop-in compatible with the pgvector extension, using the same types, operators, and query syntax, so code written for pgvector works.
Because Lakebase is a managed cloud service, a few standard-Postgres behaviors are intentionally constrained. You don't get host operating-system access or the Postgres superuser role; a databricks_superuser role takes its place, so anything that requires superuser privileges or direct file-system access won't run. Lakebase doesn't support tablespaces, and native Postgres logical replication to or from Lakebase isn't available yet. Account for these when you move an existing schema over. For details, see the Postgres compatibility reference.

FAQs

Do I need a special Databricks driver to connect?

No. Lakebase uses the standard PostgreSQL wire protocol, so standard Postgres drivers and clients connect to it. You point your existing connection string at the Lakebase host, with TLS enabled.

Which authentication methods does Lakebase support?

Native Postgres password authentication and OAuth tokens tied to a Databricks identity or service principal. Password auth suits tools and long-running processes; OAuth tokens are short-lived and rotated automatically by your application.

Can I use pgAdmin, DBeaver, or psql?

Yes. Any standard PostgreSQL client works, including psql on the command line and GUI clients like pgAdmin and DBeaver.

Does my ORM work with Lakebase?

Yes. Because Lakebase presents standard Postgres, ORMs such as SQLAlchemy, Django, and Prisma connect through their normal Postgres driver with a standard connection string.

What standard Postgres features are limited?

There is no host operating-system or Postgres superuser access, no tablespaces, and native logical replication isn't available yet. Most application code and tooling is unaffected, but review these before migrating schemas that depend on them.