Run analytics
where your data lives
DuckDB is a SQL database that runs everywhere
Now with Quack for client-server support
-- Get the top-3 busiest train stations
SELECT
station_name,
count(*) AS num_services
FROM train_services
GROUP BY ALL
ORDER BY num_services DESC
LIMIT 3;Query files and
cloud data directly
Use DuckDB's friendly but powerful SQL dialect to query any data source (Parquet, JSON, S3, data lakes, etc.)
Live demo →How we designed DuckDB
Simple
We made sure that you can install DuckDB in just a few seconds, and built it on familiar technologies so you can start using it immediately.
Read more →Feature-rich
We support a wide range of industry standard technologies (e.g., Parquet, SQL, S3 API) and combine them into a seamless user experience.
Read more →Fast
We built DuckDB on a fast columnar storage engine. DuckDB can spill to disk to support workloads far larger than the available system memory.
Read more →Free
We release DuckDB, its core extensions and the DuckLake format under MIT, a permissive open-source license.
Read more →Portable
We distribute DuckDB for all popular operating systems and ship idiomatic client APIs for major programming languages.
Read more →Extensible
We created a powerful extension mechanism that enables adding new features. In fact, we implemented many core DuckDB features as extensions.
Read more →Built for your stack
DuckDB has native clients and integrations with the data ecosystem
curl https://install.duckdb.org | sh
