Splitgraph has been acquired by EDB! Read the blog post.

Modern shouldn't mean complex.

Unify the data lifecycle with a
batteries-included modern data stack built on open protocols.

Splitgraph has allowed us to finally make sense of our pile of S3 data, as well as enabling us to easily connect downstream clients like Tableau and Metabase, which has really enabled everybody in our organisation to access our data in their most efficient way.

Harry Biddle

Data Engineer,
Stockholm Environment Institute

Many layers.
One platform.

Stop burning engineering cycles managing data infrastructure. Splitgraph combines the many layers of the modern data stack. Load, transform, discover and query data in one place.

Any to Any

Splitgraph connects numerous, unrelated data sources into a single, unified SQL interface on the Postgres wire protocol.


Experiment for a bit...

Get started quickly by querying data at its source, without ingesting it. Federated data requires little setup and makes experimenting easy.

Query live CDC data

Then commit.

Import data as versioned, immutable data "images" kept in object storage. Query any version of data by simply appending its tag to the table name.


Dozens of data sources

Import or query data from nearly 100 databases and SaaS products. Splitgraph implements the open-source Singer and Airbyte standards for data ingestion, and uses Postgres Foreign Data Wrappers for live querying. Need something custom? It's easy to extend.

illustration of sample query

Efficiency without compromise

Store your data as versioned Splitgraph images: immutable, content-addressable blocks of data in a columnar format. Inspired by Docker and Git, powered by PostgreSQL.


Separate compute and storage

Save money by decoupling storage from compute so you can scale them independently. Spin up a node pre-loaded with data, or query it immediately while Splitgraph lazily downloads the necessary objects.

Unlimited time travel

Travel back in time, or compare the same data across two points in time, by simply appending a version identifier to each table name.

Your data, your rules

You can run a Splitgraph node locally or self-host on your own infrastructure. Splitgraph Cloud is our SaaS currently in private beta, with managed and self-service options available on-prem and in all major clouds.

Map data terms to business terms

Go from raw, messy data dumps from third-party services and operational data stores to actual datasets and metrics. Start treating your data pipelines like code and easily see where every piece of data came from.


Build data images with Splitfiles

Splitfiles are a declarative way to transform data with SQL. Import from and join across any datasets and get instant provenance visualization for your data products.

Learn about Splitfiles

A home for your dbt models

Splitgraph can run your dbt models for you, giving you a unified view of your entire data pipeline. Add your dbt docs website to Splitgraph and never worry about cataloging data again.


Organize and discover knowledge

No more hunting for datasets across internal Wikis, databases and cloud accounts. Splitgraph comes with a data catalog that helps you focus on what to do with data, not how to find it.


Data on your terms


Bring your own tools

Splitgraph's Data Delivery Network is a single SQL endpoint that lets you query any version of your data with your existing tools. BI software like Metabase, data science packages like Pandas or SQL clients like DBeaver, — if it works with PostgreSQL, it works with Splitgraph.

Try now with any Postgres client

Data governance

Stop wasting weeks requesting, waiting for and granting approvals to get data. Enforce and audit access policies on your warehouse without hindering your analysts' efficiency or compromising sensitive information.


A simple web editor

Empower your non-technical team members to find and query the data they want directly from their browser.

Try it out