MotherDuck – a serverless data analytics platform powered by DuckDB

Hacker News

Share card

MotherDuck – a serverless data analytics platform powered by DuckDB

Hi folks.. Co-founder and Head of Produck at MotherDuck. We're building a new type of analytics data platform based on an exciting open source project DuckDB! We just launched today, so please head over to our website and request an invite[0]! When we first started MotherDuck a little over a year ago, we realized that the modern data platforms like BigQuery and Snowflake are of exceptional quality. They're also very mature, packing diverse set of features across a variety of use cases. So it doesn't seem like a good idea to go and compete with them... But then we found DuckDB. DuckDB changes everything - suddenly virtually any CPU in virtually any place and on any environment can be used for analytics. You can run DuckDB in the browser or a Lambda, in Python or R, in the cloud or on your laptop. Suddenly the modern cloud data warehouse looks a lot like a black-box mainframe in the cloud. So we bet big on DuckDB. We invented a new way[1] to take OSS to market - we struck an exclusive cloud partnership with the DuckDB team, and we're working in close collaboration on a common goal. We've built a serverless cloud analytics service - a place for you to store your data, query that data in SQL, share and collaborate, and leverage MotherDuck using a plethora of third party tools like dbt, HEX, and cloudquery. However, we did something very unique again - and something that's only possible with DuckDB. For any DuckDB instance out there, you can simply run '.open md:', and your DuckDB instance is suddenly supercharged by MotherDuck - you get additional capabilities like faster secure access to S3 and sharing, and you get a new type of query execution we're calling 'hybrid mode', in which your DuckDB instance is working in concert with MotherDuck-hosted DuckDB instances to execute queries in concert on data wherever it lives. In short, we intelligently route stages of queries to the right location, enabling things like seamlessly joining data on your laptop with data on S3. We've also built a web application with a catalog and a SQL IDE. We put DuckDB in our app using WASM, so we're able to cache query results and provide a low-latency interactive query results explorer so that you can sort, filter, and even pivot your data at the speed of thought. We're currently Beta, so we're slowly accepting users from the waitlist. We plan on marching towards GA and could use your help in finding out where we can improve! [0] https://motherduck.com/ [1] https://www.madrona.com/motherduck-jordan-tigani-duckdbs-hannes-muhleisen-partnerships-commercializing-open-source-projects/