Announcing autoscaling in feature-preview!Learn More

Serverless Postgres

The multi-cloud fully managed Postgres with a generous free tier. We separated storage and compute to offer autoscaling, branching, and bottomless storage.

Sign up

Made for Developers

Launch Serverless Postgres with a single command:

$ psql -h
  • Scalable

    Compute scales dynamically to ensure that you are ready for peak hours.

  • Cost Efficient

    Compute scales down to zero on usage, and cold storage offloads to S3 for cost efficiency.

  • Easy to Use

    Fully managed serverless Postgres starts in seconds.

Join the community

Learn what the experts love about Neon.

  • Gunnar Morling

    Gunnar Morling


    Neon is definitely one of the most exciting developments around #Postgres lately. The separation of storage and compute is definitely interesting, would love to see some latency numbers there. Also how "Pageservers" and "Safekeepers" are kept in sync.

  • Guillermo Rauch

    Guillermo Rauch


    With @Neondatabase , *truly* serverless PostgreSQL is finally here.You can spin up a db and connect to it in less than 3 seconds.This changes the game.

  • Leigh Marie Braswell

    Leigh Marie Braswell


    @foundersfund is so excited to partner with @neondatabase who is launching today #neonlaunch!! They offer modern, cloud-native architecture for #serverless Postgres, separating storage & compute.

  • Erik Bernhardsson

    Erik Bernhardsson


    100% Postgres compatibility and the ability to scale down to zero and cold start in less than 3s... sounds extremely cool

  • Tobias Petry

    Tobias Petry


    An *open-source* serverless PostgreSQL database is currently being built. The first one to my knowledge. 🥳 The CEO is one of the founders of @SingleStoreDB , so they really know what they do! I have big hopes for this project to succeed.

  • Johan Eliasson

    Johan Eliasson


    Ok, I just tried @neondatabase with @nhost and @HasuraHQ. Things just worked out of the box! Instant GraphQL API on a truly serverless Postgres database. The future is here!

On Demand Scalability

We separated storage and compute to make on demand scalability possible. Compute activates on an incoming connection and shuts down to save resources on inactivity.

Compute is fully client-compatible with Postgres because it is Postgres!

As the workload changes Neon adjusts the amount of resources dedicated to the compute.

Bottomless Storage

We designed our storage from the ground up as a fault tolerant scale-out system built for the cloud. It integrates with cloud object stores such as S3 to offload cold data for cost optimization. Our storage architecture ensures high availability, scale out, and unlimited capacity that we call "bottomless".

Our storage implements a "copy-on-write" technique to deliver online checkpointing, branching, and point in time restore. This eliminates expensive "size of data" backup and restore operations required for traditional database as a service systems.

Our storage technology is open source and written in Rust.

Data Branching

Neon allows to instantly branch your Postgres database to support a modern development workflow. You can create a branch for your test environments for every code deployment in your CI/CD pipeline.

Branches are virtually free and implemented using the "copy on write" technique.

Read more

Not an ordinary Postgres as a service

Neon provides true cloud native features essential for modern application development.

  • Reliable

    Neon Cloud provides high availability without a maintenance burden and a need for expert advice.

  • Auto-Backup

    Incremental auto backup functionality keeps your data safe 24/7.

  • Open Source

    Check our repository on GitHub. We develop in public under the Apache 2.0 license.

Perfect for SaaS

SaaS companies use Neon to maximize engineering velocity and minimize the cost. Our serverless architecture minimizes the cost of maintenance for inactive customers. Specifically Neon removes the need to over-provision capacity by fitting the customers into the predefined under-utilized instance sizes.

Sign up

Subscribe to Newsletter

Join us: