Last updated: September 18, 2024
# Export files to DigitalOcean Spaces with ease
**Julian Gruber**
Back-end Developer · Munich, Germany · Show bio
[](https://github.com/juliangruber)
Today, we are introducing the new [/digitalocean/store](/docs/robots/digitalocean-store.md) Robot, which allows you to export your Assembly results to DigitalOcean's[Spaces Object Storage](https://www.digitalocean.com/products/spaces/). This adds another option to the list of file exporting providers we already support, like[Amazon S3](/docs/robots/s3-store.md),[Rackspace Cloud Files](/docs/robots/cloudfiles-store.md) and[Dropbox](/docs/robots/dropbox-store.md). Check out the full list[in our docs](/docs/robots.md#file-exporting).
## The /digitalocean/store Robot
You can use the [/digitalocean/store](/docs/robots/digitalocean-store.md) Robot to upload your files or Assembly results to[DigitalOcean Spaces](https://www.digitalocean.com/products/spaces/) for long and short term storage. Your files can set to be publicly available or private to your account.
Built upon the open source [Ceph](https://ceph.io/) storage system, Spaces support most of the features [Amazon S3](https://aws.amazon.com/en/s3/) offers, which means that if you already use our[/s3/store](/docs/robots/s3-store.md) Robot, the API should look very familiar.
In this blog post, we're guiding you through the process of setting up a DigitalOcean Spaces account, transcoding your files and exporting the results to your Spaces account. If you just want to look at the Robot's API and play with it yourself,[check out its docs](/docs/robots/digitalocean-store.md) and the example Assembly Instructions below:
```json
{
"steps": {
"store": {
"robot": "/digitalocean/store",
"credentials": "spaces-store-demo",
"space": "spaces-store-demo",
"acl": "public-read"
}
}
}
```
## Setting up DigitalOcean Spaces
The first step is to create a DigitalOcean account, if you haven't already done so. Head over to[digitalocean.com](https://www.digitalocean.com/) to get started.
Next, navigate to the **Spaces** tab to create a new Space, which we're going to call`spaces-store-demo`. Select the datacenter region closest to where you're going to fetch the data from. Here, we've selected `fra1` / Frankfurt.
Once created, we have an empty Spaces account to which we can start exporting files:
In order for our Robot to be able to talk to the[DigitalOcean Spaces API](https://developers.digitalocean.com/documentation/spaces/), we also need to create a pair of Spaces access keys. You can do so in the **API** section of your DigitalOcean account. Best keep this tab open for now, as we're going to need the public and private secret keys in the following steps. Never share these credentials with anyone you don't want to have write access to your Space.
## Creating Template Credentials
We're done with DigitalOcean for now, so let's head over to[transloadit.com](/index.md) and create a free account! The Sandbox plan comes with 2 GB of transcoding volume and two priority job slots, so it's all you're going to need to get started with our Robots.
To allow better security and easier re-use, Transloadit offers you to create Template Credentials and store the secrets for you. Once logged in, navigate to your account and select**Template Credentials**. Choose the third party service `Digital Ocean Spaces` and enter your Space credentials as follows, replacing *space name*, *region*, *key id* and *secret access key* as required:
Hit save and you're done! The Template Credentials name — `spaces-store-demo` in this case — is what we're going to use in the next step.
## Detecting faces
Let's get to finally submitting some files to a new Assembly! Our example is going to use the [/image/facedetect](/docs/robots/image-facedetect.md) and[/digitalocean/store](/docs/robots/digitalocean-store.md) Robots. Once we're done, you'll be able to upload image files in which all human faces are going to be detected and separately uploaded to your DigitalOcean Space.
For the purpose of this guide, we're using the [curl(1)](https://curl.haxx.se/) command line tool to interact with the Transloadit API, which is most likely already installed on your computer. If you are on Windows, check out the official[curl for Windows instructions](https://curl.haxx.se/windows/).
First, another preparatory step: head over to the [API settings](/c/settings/) in your Transloadit account and please copy the **Auth Key**.
(Don't worry, we regenerated our Secret after taking this screenshot 😄)
Then, create a file with these Assembly Instructions, which we're going to call`params.json`. Replace the `auth.key` with your personal **Auth Key**.
```json
{
"auth": {
"key": "3ba3ac15473f4ce89a279db977248d29"
},
"steps": {
"face-detect": {
"use": ":original",
"robot": "/image/facedetect",
"crop": true,
"faces": "each"
},
"store": {
"use": "face-detect",
"robot": "/digitalocean/store",
"credentials": "spaces-store-demo",
"space": "spaces-store-demo",
"acl": "public-read"
}
}
}
```
The Assembly consists of two Steps:
1. Detect faces and extract each detected face as a cropped image
2. Upload the extracted faces to your DigitalOcean Space, using `public-read` permissions, so the files are publicly accessible
## Testing the Assembly
Now there is just one piece missing, and that's to kick off the whole machinery using a single command!
You're going to need an image with some faces on it to test your Assembly. Here's one you can grab if you don't have one readily available:
Issue the following `curl(1)` command in your terminal, in the same folder where `params.json` and the source image (here `group.jpg`) are stored. Pipe to[json\_pp(1)](https://metacpan.org/pod/distribution/JSON-PP/bin/json%5Fpp) in order to get a nicely formatted JSON output.
```bash
curl \
-s \
-XPOST \
-F params="$(< params.json)" \
-F original=@group.jpg \
https://api2.transloadit.com/assemblies \
| json_pp
```
> If you're curious why the `curl(1)` command looks as it does:
>
> * `-s` instructs `curl(1)` to be quiet, so we only see the HTTP response body
* `-XPOST` in order to make this a POST request
* `-F ...` Transloadit uses[HTTP multipart formposts](https://everything.curl.dev/http/post/multipart), so that the client can submit both the Assembly Instructions and the files, all in one request (and so that our API can be integrated with plain old `