Built by people who love data
We've spent a long time around databases, from the character-mode era to today, and we still get a quiet thrill from a clean schema and a query that does exactly what it should. SnoutData is everything we wanted around that work, in one place: a workbench on your machine and a backend in the cloud.
What SnoutData is
SnoutData covers the whole of working with data: the databases you already run, the data that lives around them, the agents that now work beside you, and a backend for the app you're building. It is two products, one on your machine and one in the cloud, under one account.
SnoutData Studio is the workbench on your machine: a SQL editor that knows your schema, Data Flows that read files, PDFs, web pages, cloud buckets and logs into the destination you choose, moving a database into the cloud and back, generating datasets with a model you pick, monitors that watch your database, a 3D map of your schema, and a dock where your own coding agents (Claude Code, Codex, opencode) work next to you.
SnoutData Cloud is the backend. Every project is a Postgres database with auth, storage, realtime, functions and push notifications, and time series built in. And if you'd rather keep the backend on your own hardware, you can download our cloud stack, run it locally in Docker, and manage it from SnoutData Studio.
Two products, two promises
The two keep your secrets in opposite ways, on purpose.
SnoutData Studio runs on your machine: your credentials stay in your system keychain, and the agents you run in it never receive them.
SnoutData Cloud is a hosted backend, and a host holds what it hosts. We keep your project's credentials encrypted, you can view or reset them whenever you like, and every change is on an audit log. If you want credentials that never leave your machine, that's what Studio is for, or run the stack yourself. We'd rather say so than pretend otherwise.
Why a smaller model is enough
Most everyday data work isn't exotic. It's finding the right table, remembering how two tables join, writing a correct query, and not breaking production. Reaching for the most expensive model on the planet to write a simple, schema-correct query is like renting a crane to hang a picture frame.
So we prepare instead. SnoutData indexes your schema on your machine and hands the model exactly the tables, columns and relationships your question touches. It understands the app itself, not just your data, so answers land where you're working. With that context, a smaller model gives grounded, column-correct answers. Same outcome, a fraction of the cost and fuss.
What's open, and what isn't
We grew up on open source. The languages, databases, editors and libraries we've relied on for years were given freely by people we'll never meet, and SnoutData stands on a tall stack of that generosity. SnoutData Studio itself is closed source. That's what lets an independent project stay funded and keep improving for the long run.
A lot of the rest is open, under Apache-2.0 on GitHub: the snoutdata CLI, the @snoutdata/client library, SnoutTime (the time-series extension in every Cloud project), and a growing set of the services and Postgres extensions Cloud runs on. If you think we got something wrong, tell us.
Come say hello
We'd like to hear from you. Download the app or start a Cloud project and give it a real try, follow along on GitHub, or email us at [email protected]. Thanks for reading this far, and thanks for caring about getting your data right.
The SnoutData team, Toronto, Ontario, Canada