Databricks Feature Store
docs.databricks.com
- Free trial
- Web
- API
What Databricks Feature Store is
Databricks Feature Store is a central registry for features used in AI and machine-learning models. Registering features and models in Unity Catalog provides governance, lineage, point-in-time joins, and discovery and sharing across workspaces. Teams can author features as declarative Feature Views managed through Databricks pipelines or as feature tables populated in Delta tables. Feature Views and their materialization capability are in Public Preview. Models trained with Feature Store features track feature lineage and look up current feature values at inference. Databricks says using the same feature computations for training and inference eliminates training/serving skew. Feature-serving endpoints and online stores support real-time applications, with the overview describing millisecond latency. Third-party online-store options include Amazon DynamoDB, Amazon Aurora (MySQL-compatible) and Amazon RDS MySQL, with availability varying by mode. The paid cloud service has a free trial and usage-based monthly billing. It requires a Unity Catalog-enabled workspace, and its APIs cannot be called from local or other non-Databricks environments, though local IDE development and unit testing are supported.
Who it suits
It suits teams building AI or ML models in Databricks that need feature governance, lineage and serving. It requires a Unity Catalog-enabled workspace.
Good to know, and what to watch
- Provides governance and lineage through Unity Catalog
- Supports point-in-time joins and cross-workspace discovery
- Tracks feature lineage for trained models
- Supports online stores and real-time serving
- Free trial is available
- Requires a Unity Catalog-enabled workspace
- APIs cannot be called from non-Databricks environments
- Feature Views are in Public Preview
Plans
Every plan and pricePay as you go
Getting started
- Open the Databricks Feature Store documentation website
- Use a Databricks workspace enabled for Unity Catalog
- Choose the paid service and its free trial
- Author Feature Views through Databricks pipelines or populate Delta-backed feature tables
- Install the databricks-feature-engineering Python client in Databricks Runtime 13.3 LTS ML or newer
- Configure online or offline stores and serving endpoints as needed
Limits to check
The service requires Unity Catalog, and Feature Engineering and Feature Store APIs cannot be called from local or other non-Databricks environments. Feature Views and their materialization capability remain in Public Preview.
The verdict
Databricks Feature Store connects feature management with model training and inference in a Unity Catalog workspace. Teams should account for its workspace requirement, API limitation and usage-based pricing components.
Questions people ask
Is Databricks Feature Store free?
It is a paid service with a free trial. Pay-as-you-go billing is monthly and based on usage.
Which platforms does it support?
It provides web and API access as a cloud deployment.
Does it support real-time serving?
Yes. Feature-serving endpoints and online stores support real-time applications, with the overview describing millisecond latency.
Which online stores integrate with it?
Supported third-party stores include Amazon DynamoDB, Amazon Aurora (MySQL-compatible), and Amazon RDS MySQL, with availability varying by Feature Store mode.
Can it run from a local environment?
Its Feature Engineering and Feature Store APIs cannot be called from local or other non-Databricks environments, although local IDE development and unit testing are supported.
What the maker publishes
- Client limitation
- Feature Engineering and Feature Store APIs cannot be called from local or other non-Databricks environments, although local IDE development and unit testing are supported. source
- Feature authoring
- Features can be authored as declarative Feature Views managed through Databricks pipelines or as feature tables populated by writing values to Delta tables. source
- Feature Views status
- Feature Views and their feature materialization capability are in Public Preview. source
- Governance and discovery
- Registering features and models in Unity Catalog provides governance, lineage, point-in-time joins, and cross-workspace feature sharing and discovery. source
- Headquarters
- San Francisco, California, United States source
- Integrations
- Supported third-party online stores include Amazon DynamoDB, Amazon Aurora (MySQL-compatible), and Amazon RDS MySQL, with availability differing by Feature Store mode. source
- Model training and inference
- Models trained with Feature Store features automatically track feature lineage and look up the latest feature values at inference time. source
- Pricing components
- Materialization uses serverless compute, feature serving endpoints use the Model Serving SKU, and online stores use Lakebase compute priced by capacity units and replicas. source
- Pricing model
- Feature Store is billed at cost for underlying serverless compute, online store, and serving infrastructure, with no premium on top. source
- Purpose
- Databricks Feature Store is a central registry for the features used in AI and ML models. source
- Python client
- The Feature Engineering Python client is distributed as databricks-feature-engineering on PyPI and is pre-installed in Databricks Runtime 13.3 LTS ML and above. source
- Real-time serving
- Feature serving endpoints and online stores support serving features for real-time applications, with the main overview describing endpoints as providing millisecond latency. source
- Security and access
- Access to Unity Catalog feature tables is managed through Unity Catalog access controls. source
- Training and serving consistency
- Databricks says using the same feature computations at inference as during training eliminates training/serving skew. source
- Workspace requirement
- The current Databricks Feature Store requires a workspace enabled for Unity Catalog. source
Checked against the maker's pages on 30 Sept 2026.