Integrations

Connect Lakebase to the rest of Azure Databricks: register databases in Unity Catalog for governance, serve lakehouse data through Postgres for low-latency reads, feed Postgres changes back to the lakehouse, and use Lakebase as a backend for AI agents, feature serving, and search. For the architecture behind these integrations, see LTAP architecture.

Integration	Description
Register database in Unity Catalog	Register your Lakebase database in Unity Catalog for unified governance and cross-source queries.
Serve data with synced tables	Sync Unity Catalog table data into Lakebase to serve it to applications at low latency.
Lakebase Change Data Feed	Store row-level changes from Lakebase Postgres tables as Unity Catalog Delta tables for downstream pipelines, audit, and external consumers.
Agent state and memory	Persist agent state and memory in Lakebase so AI agents built with LangGraph or the OpenAI Agents SDK retain context across sessions.
Feature Store and Model Serving	Serve ML features at low latency from Lakebase for models deployed with Databricks Model Serving.
Lakebase Search	Add hybrid vector and keyword search to your project using the lakebase_vector and lakebase_text extensions.

Additional resources

• LTAP architecture
• Connect to your database
• Query your data