---
title: Built on Bun
description: Bun runs your server code, bundles your pages and installs your packages. The data stays Sluurp's.
section: Get started
order: 4
---

# Built on Bun

Bun runs your server code, bundles your pages and installs your packages. The data stays Sluurp's.

Sluurp is one executable with [Bun](https://bun.sh) inside it. Bun does what it does best: running JavaScript, bundling it and installing packages. Sluurp does the rest: the database, rules, sign-in, sync, history and the admin.

## What Bun does

- **Runs your server code.** Functions, routes, jobs and queues run in Bun's JavaScript engine, each in its own worker. Everything a function asks of Sluurp for the person calling it, such as `ctx.collection("posts").list()`, still goes through Sluurp's rules.
- **Bundles your pages.** Bun bundles React apps and pages, and updates them as you save.
- **Installs your packages.** `sluurp bun add zod` runs Bun's package manager, and `sluurp bunx` its package runner. Packages that need Node's APIs, such as database drivers, work in your server code.

## Answer a request yourself

A route that exports a handler for a method answers that method itself:

```ts title="routes/visits.ts"
import { SQL } from "bun"

const pg = new SQL(process.env.DATABASE_URL!)

export async function GET() {
  return Response.json(await pg`select count(*)::int as visits from visits`)
}
```

This route uses Bun's own Postgres client. Bun also speaks MySQL and SQLite, and drivers from npm work too, such as `oracledb` for Oracle.

## Your data

The database is Sluurp's SQLite, with full-text search and [vector search](/docs/search) built in. Your code reaches it through `ctx`, so collection rules apply to everything a function does for the person calling it.

Bun's own SQLite (`bun:sqlite`) can open the same file read-only, for a report or an export. Vector search isn't available there: search by meaning through `ctx`, with `near`.
