Spiral Platform
Build data applications on Spiral: sign-in, permissions, versioned tables, commit history, background jobs, and vector indexes.
The Website Terms of Use govern standalone use of these docs.
Spiral is a platform for building data applications. It supplies the parts that every data application needs: sign-in and permissions for your users, versioned tables in object storage you control, a commit history for every change, background jobs, and vector indexes. You write the application. Spiral runs the services, and your code reads and writes the data directly in storage.
Spiral builds its own applications on the same platform and hosts them for you to use. Spiral Robotics imports robot episodes, curates them and trains policies on them. Spiral Samples curates the rows of a Table into training sets. Rerun on Spiral opens recordings stored in Spiral in the Rerun viewer. These apps are in development, and so is an OpenTelemetry app. See Example Apps.
What works today
- HTTP APIs for every service, at
https://api.spiraldb.dev: Organizations, Projects, Authentication, Authorization, Workloads, Applications, the Catalog, Tables, Enrichments, Vector Indexes, Locations, Connections, Storage, Tasks, and Worker Pools. Some of them, such as Applications and Tasks, are still in development, and their pages say so. - The Rust SDK, published as
spiral-sdkon crates.io, manages resources, publishes commits, and reads and writes Tables in your own process. - The Python SDK, published as
spiral-sdkon PyPI, and the TypeScript SDK, published as@spiraldb/sdkon npm, call service APIs and read and write Tables. - The DataFusion integration, so you can query Tables with SQL in your application's own DataFusion session.
- Tools: the
spiralCLI, SQL and local Connect from the Python SDK, Spiral Studio, and a read-only MCP server for AI agents.
Planned
- A Spiral-hosted data plane, for browsers and SDKs without the native library, and a data plane that you run in your own cloud.
- Polars and DuckDB integrations.
Start here
Build the smallest Spiral application: create a Table from Rust and read it back.
See what Spiral runs, what your code runs, and why bytes never pass through Spiral.
Where data goes, how Projects are kept apart, and how long each token lives.
Learn the model
The account that holds your people and the workspaces that hold your data.
Tables and the other things you store, their versions, and their addresses.
Change several Tables at once and pin the exact data a job read.
Give people and workloads roles on a Project.
How these docs are organized
The docs have five tabs. Platform, this tab, explains the model that every service shares, the data plane that reads and writes your data, and the conventions of the HTTP API. Apps describes the applications we build on Spiral. Services has one section per service, such as Tables or the Catalog, with its concepts, task guides and HTTP reference. SDKs & Tools covers the Rust SDK, the integrations, the CLI and Studio. Reference lists permissions, error codes, resource addresses and SPQL functions.