Hydra adds the open‑source pg_duckdb extension to PostgreSQL, embedding DuckDB’s columnar engine to store analytics tables in a compressed columnstore and run queries with serverless, autoscaling compute. The platform delivers sub‑second query latency, automatic caching, and bottomless storage while isolating analytics workloads from OLTP, and integrates with standard SQL, ORMs, BI tools, and cloud object stores via a unified console and API.
Funding
Funding not disclosed
Founders
Product
Problem
Organizations that need to run analytical queries on their transactional PostgreSQL data must maintain separate data warehouses or ETL pipelines, which adds latency, operational overhead, and cost while limiting real‑time insight.
Solution
Hydra extends PostgreSQL with the open‑source pg_duckdb extension, embedding DuckDB’s columnar engine directly inside the database. It automatically stores analytics tables in a compressed columnstore and isolates compute in a serverless layer that scales per query, delivering sub‑second response times without impacting OLTP workloads. Automatic caching and bottom‑less storage eliminate the need for manual data movement or capacity planning. The platform integrates with existing ORMs, BI tools, and cloud object stores, letting developers query row‑store and column‑store tables with standard SQL. All management functions—scaling, monitoring, backups, and role‑based access—are provided through a unified web console and API, enabling teams to add real‑time analytics to any PostgreSQL stack instantly.
Target Audience
Hydra is aimed at data‑engineers, backend developers, and product teams building real‑time analytics on event, time‑series, or IoT data within PostgreSQL, as well as SaaS and fintech companies that require low‑latency BI without separate warehouses.
Features
- PostgreSQL‑native pg_duckdb extension that runs DuckDB’s vectorized execution engine inside the database.
- Serverless compute autoscaling per query, with isolated resources to avoid contention with transactional workloads.
- Automatic caching layer that guarantees predictable sub‑second query latency at any scale.
- Bottomless columnar storage with 10×‑15× data compression and automatic growth, eliminating storage‑capacity planning.
- Seamless JOIN support between row‑store (heap) and column‑store tables using standard SQL syntax.
- Full compatibility with popular ORMs and drivers (node‑postgres, drizzle‑orm, postgres.js) and BI connectors.
- Integrated monitoring, scaling controls, snapshot/WAL backups, and multi‑region deployment via a web dashboard and REST API.
- Role‑based access control and end‑to‑end TLS encryption for secure multi‑tenant environments.