Skip to main content

Feldera runs arbitrary SQL queries that instantly update when new data enters your lakehouse. Experience instantly computing fresh features and enhanced accuracy for your ML pipelines, without waiting hours for your complex batch and ETL pipelines to complete. Get results orders of magnitude faster and keep your data teams lean.


Key Features

Feldera is powered by a state-of-the-art, open-source streaming engine we built from the ground up. It lets users express their analytics and transformations as SQL queries, which the platform continuously evaluates automatically, incrementally, correctly and efficiently.


Real-time SQL without compromises

Feldera does not limit the kinds of analytics you can do on streaming data. 1) If we can't run a SQL program, that's a bug. 2) Feldera seamlessly accepts all change events (insertions, updates and deletes) from multiple sources. 3) Join, transform, analyze and aggregate the data, and send the outputs as changes to multiple sinks.


Accelerate time-to-result in your lakehouse

Feldera integrates with your lakehouse ecosystem seamlessly. It can maintain views in your lakehouse using open-formats (like DeltaTable), allowing you to query them from your existing ecosystem of choice (like Databricks) as normal tables.


Built on a foundation you can trust

All the above is uniquely possible because we based our streaming engine on a solid foundation, DBSP, our award-winning research on automatic incremental view maintenance. With our team's combined decades of research experience on distributed systems, databases and operating systems, we are on a mission to bring the state of practice to match the state of the art.