Skip to main content
Write with AI. Run locally. Deploy to Xano. This section is the middle step: your coding agent writes the backend in TypeScript, and you run it on Xano itself before anything reaches your workspace. There are two places to run it, and they are the same backend: Deploy to the Engine while you build:
On Windows or an Intel Mac? There is no Xano Engine build for your machine yet. Wherever this page says npm run xano:deploy, run npm run xano:deploy:ephemeral instead: it deploys the same code to an ephemeral in Xano’s cloud. That needs a Xano account, so sign in first (Connect to Xano). See where the Engine differs today.
The first deploy downloads the Engine and pins its version in package.json. Every later one merges your changes into the running Engine and keeps the rows you entered. A few gaps between the Engine and an ephemeral change what you run today; they are listed once, in Where the Engine differs today.

​
In this section

The Xano Engine

Deploy to Xano on your machine, pin its version, keep rows, and manage engines.

Engine MCP server

Let Claude Code, Cursor or Codex read rows and run functions and tests on the Engine.

Ephemerals

The same backend in Xano’s cloud, with its own URL and an expiry.

Frontend hosting

Serve your built frontend beside the backend, and render SvelteKit on the server.

​
Next

When a build is ready to ship, deploy it to an ephemeral and cut a release: Deploy to Xano.