Working with Collections

A collection in Estuary can be thought of as a bucket of CIDs put together. Collections are used to organize files and other types of data uploaded to Estuary. There is no restriction about the type of data that can be held in a collection, provided it's pinned on, or in other words, uploaded to Estuary, and that it has a CID.

Creating a collection

Before being able to use a collection, it needs to be created. Here's how to do it using curl:

curl -X POST -d '{ "name": "My super nice new collection", "description": "This collection holds only the best content" }' -H "Content-Type: application/json" -H "Authorization: Bearer REPLACE_ME_WITH_API_KEY"

Obs: If you're running your own Estuary node, you will need to change to your own address such as https://localhost:3004

Upon success, Estuary returns the created collection:

  "createdAt": "2022-02-23T16:20:50.518337618-03:00",
  "uuid": "28d923b5-2561-43ee-8ab3-fb42088666f2",
  "name": "My super nice new collection",
  "description": "this collection holds only the best content",
  "userId": 1

We're specially interested in the uuid field here, which is the unique identifier we will use to add data to that collection in the following section.

Adding a file to a collection

There are multiple ways to add data to a collection. Here we will explore two endpoints: /collections/add-content and /content/add-ipfs.


Let's add two CIDs QmW3sbi25Veqkg3o9qCMkTuosSw6S8hSPzAEffwA1tCf5S and QmS8dypUY34t3UF7Xd98KhuxqQ8F45WckJCGkdhNnwgvM4 using this endpoint:

curl -X POST -d '{ "contents": [], "cids": [QmS8dypUY34t3UF7Xd98KhuxqQ8F45WckJCGkdhNnwgvM4, QmW3sbi25Veqkg3o9qCMkTuosSw6S8hSPzAEffwA1tCf5S], "collection": "28d923b5-2561-43ee-8ab3-fb42088666f2" }' -H "Content-Type: application/json" -H "Authorization: Bearer REPLACE_ME_WITH_API_KEY"

Obs: Notice we used the uuid of the collection we created earlier (28d923b5-2561-43ee-8ab3-fb42088666f2) to identify the collection we want to upload content to.

We can also use estuaryIds to specify contents that are already uploaded to Estuary in order to include them in a collection. When a file gets uploaded to Estuary, it gets a corresponding estuaryId. Let's first add a file (file1.txt) to estuary using the /content/add endpoint in order to get its contentID:

curl -X POST -H "Authorization: Bearer REPLACE_ME_WITH_API_KEY" -H "Accept: application/json" -H "Content-Type: multipart/form-data" -F "data=@/tmp/file1.txt"

  "cid": "bafkqadlumvzxi2lom4qgm2lmmufa",
  "estuaryId": 10,
  "providers": [

Obs: estuaryId can also be called contentID or contID

Now that we have the estuaryId (which is 10 in this example), we can add file1.txt to our collection, again using its UUID to identify it:

curl -X POST -d '{ "contents": [10], "cids": [], "collection": "28d923b5-2561-43ee-8ab3-fb42088666f2" }' -H "Content-Type: application/json" -H "Authorization: Bearer REPLACE_ME_WITH_API_KEY"


We can also use the /content/add-ipfs endpoint to add CIDs to estuary and also put them in a collection at the same API call. Using this endpoint we must specify the content by CID (in the root field):

curl -X POST -d '{ "name": "file1.txt", "root": "QmS8dypUY34t3UF7Xd98KhuxqQ8F45WckJCGkdhNnwgvM4", "collection": "28d923b5-2561-43ee-8ab3-fb42088666f2", "collectionPath": "/dir1/file1.txt" }' -H "Content-Type: application/json" -H "Authorization: Bearer REPLACE_ME_WITH_API_KEY"

Notice we also specified a new field collectionPath. In the next section we will explore another feature of collections: collection directory paths.

Collection directory paths

Besides having several collections to organize data, users can also further organize content inside collections using directory paths. Directory paths are filesystem-like paths such as /this/is/a/path/to/a/file. To create a path, we just need to put a file inside it using the collectionPath field. Let's take the example used in the last section:

curl -X POST -d '{ "name": "file1.txt", "root": "QmS8dypUY34t3UF7Xd98KhuxqQ8F45WckJCGkdhNnwgvM4", "collection": "28d923b5-2561-43ee-8ab3-fb42088666f2", "collectionPath": "/dir1/file1.txt" }' -H "Content-Type: application/json" -H "Authorization: Bearer REPLACE_ME_WITH_API_KEY"

Since the collectionPath for file1.txt is /dir1/file1.txt, the path /dir1/ inside that collection gets created, and we can list only the contents of that path:

curl -X GET -H "Authorization: Bearer REPLACE_ME_WITH_API_KEY" ""
    "name": "file1.txt",
    "dir": false,
    "size": 20,
    "contId": 11,
    "cid": "QmS8dypUY34t3UF7Xd98KhuxqQ8F45WckJCGkdhNnwgvM4"

Listing all the contents of a collections

When a piece of content doesn't have a directory path in a collection, it is impossible to list it using the /collections/fs/list endpoint since that requires a dir to be specified and, even if it isn't, it defaults to the / path. In order to list all the contents of a collection, even the ones that don't have a directory path set, we can use /collections/content/:collection-uuid:

curl -X GET -H "Authorization: Bearer REPLACE_ME_WITH_API_KEY"