SDKs & Tools
Choose the client library, integration, or tool that your application and team use with Spiral.
Your application reads and writes Spiral data through a client library, and the people on your team look at the same data through Studio, the CLI, or an AI assistant. Every client works with the same Projects and Tables, so rows that your Rust service writes are immediately visible to everyone else without an export step.
If you are starting an application today, use the Rust SDK. It covers every service, and the Python and TypeScript SDKs are built on the same Table and commit model. Add the DataFusion integration if your application needs SQL.
Client libraries
| Client | Status | Use it to |
|---|---|---|
| Rust SDK | Available on crates.io. | Manage Organizations, Projects, Workloads, Applications and installations; read and write Tables; create and refresh Enrichments; run SPQL functions on Arrow batches. |
| Python SDK | Available on PyPI. | Call every service's API, read and write Tables, publish commits, create and refresh Enrichments, and run SPQL functions from Python. |
| TypeScript SDK | Available on npm. | Call every service's API, read and write Tables, and show authenticated native media with browser cookie connections or Node bearer connections. Experimental browser workers execute modern Table reads and writes with OPFS storage. |
Integrations
| Integration | Status | Use it to |
|---|---|---|
| DataFusion | Available on request. Not yet on crates.io. | Query Tables with SQL in your application's own DataFusion session. |
| ADBC | In development. | Query Tables from an ADBC client, including Python DBAPI through the separate spiral-sql package. |
| PyTorch | Included in the published Python SDK; install spiral-sdk[torch]. | Feed exact Table versions to a training loop. |
Tools
| Tool | Use it to |
|---|---|
| CLI | Install spiral-cli to sign in and manage resources. Add spiral-sql for terminal SQL through spiral sql and local Connect through spiral connect. Both packages are available on PyPI. |
| Harlequin | Query Tables through the Spiral ADBC adapter in an interactive SQL terminal. |
| Studio | Browse Projects and Table data; configure table compaction, workload identities, credentials, and Project roles in the browser or desktop app; run Tasks on your own machine in the desktop app. |
| MCP server | Let an AI assistant such as Claude list your Projects, branches, and resources. |
The Python SDK, spiral-cli, and spiral-sql have independent versions and
releases. Install or upgrade each package for the capabilities your application needs.
Without a client library
Every service also has an HTTP API. Use it from a language that has no SDK yet, or for operations your client does not expose. Reading and writing Table rows needs an SDK, because the SDK moves the bytes directly between your process and storage.