High-Performance PostgreSQL
Postgres, crafted for the modern era.
We took everything that makes PostgreSQL great and rethought the parts that needed it — a modern async architecture, our own accelerated query execution engine, and enhanced planner. Still PostgreSQL, meaningfully faster & more scalable.
Under the hood
Three fundamental improvements to the PostgreSQL architecture. Click to explore.
Vanilla PostgreSQL spawns a separate OS process for every connection, making connection pooling an external necessity and limiting the potential for intra-query parallelism.
We've upgraded this. Our async dispatch engine means:
- No need for external poolers — PgBouncer is no longer necessary.
- Parallel query workers are extremely cheap to spawn, unlocking real concurrency.
- New capabilities like USE database for seamless in-session database switching, opening up flexible multitenancy patterns.
PostgreSQL's executor is solid, but there are workloads where modern techniques can deliver dramatic improvements:
- Aggregations
- String operations — full-text search, LIKE/regex
- JSON processing
- Large sequential and index scans
- Index merge operations
Our executor uses SIMD instructions, compressed bitmaps, and targeted algorithmic optimizations for specific query shapes. Combined with our JIT compiler, most queries run delivering up to 25x speedup on targeted workloads and an average of ~2.5x on regular query mixes.
PostgreSQL's planner was designed ~30 years ago under assumptions of limited memory and storage. As a result, it doesn't collect enough statistics under any configuration to make truly optimal plans at scale.
We built a modern streaming statistics collector and introduced new kinds of targeted statistics that give the planner a much richer picture of your data. The result: more efficient execution plans, especially as your dataset grows.
The larger, the more complex your data, the bigger the impact — queries that used to degrade with scale now stay fast.
Built by people who know the internals
Our team has deep roots in database engineering and a breadth that goes well beyond it — SaaS platforms, low-level systems architecture, programming language runtimes, networking, and more. We've spent years building and operating complex infrastructure at scale.
More about the team →The PostgreSQL you know. The performance you need.
We're working with early design partners now. If you're running PostgreSQL at scale and want to push the limits, let's talk.