Snowflake App Runtime adds serverless compute and Python support
Snowflake is expanding App Runtime with serverless compute, browser-based development in Workspaces and Python support, giving organizations more ways to build applications alongside data already stored in Snowflake. The features are in public preview.
- Serverless apps suspend when idle and resume in seconds.
- Workspaces supports building and deploying Next.js apps in a browser.
- Python 3.13 support includes FastAPI and Flask deployments.
- An admin page provides a view of apps across an account.
- Horizontal scaling for serverless apps is in private preview.
Apps can suspend when idle
Snowflake App Runtime has been generally available since September. The service is designed to let organizations build internal tools and agentic applications within their Snowflake accounts, with built-in sign-in and access to Snowflake governance controls.
With the new serverless compute option, apps automatically suspend when they become idle and resume in seconds when someone opens them. According to Snowflake, customers pay for compute while an app is running, rather than keeping compute active between uses. The model is intended for tools with different usage patterns, such as morning dashboards, weekly approval applications and month-end reporting apps.
Each serverless app currently runs as a single instance. Snowflake recommends testing heavily used applications under realistic load before moving them to serverless. Horizontal scaling, including a minimum number of warm instances, is in private preview. Serverless apps also support custom URL prefixes, with addresses that remain unchanged across redeployments.
Workspaces brings app development into the browser
App Runtime support in Snowflake Workspaces is in public preview for Next.js applications. Developers can open a browser, describe an app to Snowflake CoCo, view it in a live preview panel and deploy it without setting up a local development environment or additional configuration.
Snowflake says the workflow can take an application from its first prompt to a live URL within the customer’s Snowflake account. There is no compute setup for preview, and deployed apps operate under the account’s governance controls because they run inside Snowflake.
The company is also introducing Python support for serverless App Runtime. Developers can bring frameworks including FastAPI and Flask, with dependencies installed through pip or uv. Python applications run on Python 3.13 and receive built-in sign-in, Snowflake governance controls and custom URLs.
Administration tools target growing app portfolios
Snowflake is adding an App Runtime administration page that brings every app in an account into one place. Administrators can filter applications by status and compute type without writing a query.
For Node.js developers, the App Runtime SDK, @snowflake/sar-sdk on npm, provides documented patterns for sign-in, owner’s and caller’s rights queries, and secrets. Snowflake says CoCo’s App Runtime templates use the SDK, while other coding agents can follow the same patterns.
The additions could make App Runtime more practical for organizations that want teams to replace spreadsheets, manual handoffs or repeated data requests with internal applications. Snowflake says managed compute will remain available for now. When serverless becomes generally available, the company plans to make it the default compute model and end the managed compute discount, with timing and details to be shared before any changes.
Snowflake: Snowflake App Runtime Goes Serverless, Adds Python Support


