---
title: Cloudflare R2
description: Store artifacts in an R2 bucket bound to the Worker, and choose the bucket's name and location.
---

The Worker stores each artifact as one object in the bucket bound as `R2_STORE`. This is the default configuration.

## Use another bucket

`apps/api/wrangler.jsonc` binds a bucket named `turborepo-cache`:

```jsonc apps/api/wrangler.jsonc
{
    "r2_buckets": [
        {
            "binding": "R2_STORE",
            "bucket_name": "turborepo-cache",
            "preview_bucket_name": "turborepo-cache-preview",
        },
    ],
}
```

To use another bucket, change `bucket_name` and deploy again. If the bucket doesn't exist, `wrangler deploy` creates it. Keep the binding name `R2_STORE`, because the server looks for that name.

## Choose where the bucket lives

Without a location hint, R2 creates a bucket near the request that creates it. When Wrangler creates the bucket, that request comes from the machine or CI runner that runs the first deploy.

To pick the location, create the bucket before the first deploy. Choose the hint closest to most of your builds, such as your CI region:

```sh
cd apps/api
pnpm exec wrangler r2 bucket create turborepo-cache --location enam
```

Cloudflare lists the hints in [Data location](https://developers.cloudflare.com/r2/reference/data-location/#location-hints). R2 sets the location only when it creates a bucket. To move the cache, create a bucket in the new location and change `bucket_name`.

## Use R2 from another platform

A server outside Cloudflare reaches R2 through its S3-compatible API. [S3-compatible storage](/storage/s3#cloudflare-r2) shows the variables.