It’s our mission to make the best all-in-one database for application development so every developer can ship apps, analytics, and ML on realtime data in an afternoon.

With Hydra, analytical processing runs directly on live transactions and events with zero-ETL. Every Hydra project includes a fully managed Postgres database with sub-second analytical processing. We built Hydra on a bottomless, globally distributed object storage so you can scale from MVP to enterprise scale.

Hydra’s design is simple: we put the world’s fastest growing analytics engine, DuckDB, into the world’s most advanced open source database, Postgres. Traditionally, event-driven applications require a transactional database, an analytical database, an S3 bucket, and data pipelines. This approach relies on ETL pipelines between Postgres and an external analytics DBMS. By removing the friction of ingesting and transforming data between a transactional rowstore and analytical columnstore, Hydra operates 100-1000X faster than traditional architectures.

Most analytics DBMS (OLAP) today are designed for data engineers, operations teams, and data scientists: That has to change if we want richer, faster, applications built by software engineers. By leveraging DuckDB‘s “sqlite for analytics,” capabilities every developer can transform standard Postgres into the all-in-one database we’ve all dreamed of. The future of data is realtime, intelligent, and ducking-awesome!

​
Hydra product non-negotiables

  • Best analytics performance & features on earth

  • Best transactional performance & features on earth

  • Set up in less than a day

    • No learning: it’s Postgres.

    • Remove ETL pipeline setup and costs

    • No application code or major schema changes

    • Zero-downtime data migration

  • Open source (MIT) with zero external dependencies

  • Fully managed cloud database

    • Custom-built Postgres with super-user access

    • Backups, Disaster Recovery, High Availability (HA)

    • Bottomless, globally distributed object storage

    • isolated, multi-user concurrency

    • Advanced data compression

pg_duckdb (open source)Architecture