List the items of a collection

Returns the published documents of one collection, such as legal-pages, translations or release-notes, as a page of results. Your documents and Ovok's shared ones are merged.

Method	Path
`GET`	`/v1/public/cms/collections/:collection/items`

:collection is the collection's slug. Collection slugs are case-sensitive, and the users and tenants collections are never served.

Who can call it

An open collection (legal-pages and translations by default) needs only a tenant. Any other collection needs a CMS API key or a signed-in user of the tenant. See who can read content.

Request

curl --get 'https://api.sandbox.ovok.com/v1/public/cms/collections/legal-pages/items' \
  --data-urlencode 'tenant=big-health-company' \
  --data-urlencode 'locale=de'

A collection that is not open takes a key, from a server:

curl --get 'https://api.sandbox.ovok.com/v1/public/cms/collections/release-notes/items' \
  --header "x-api-key: ${OVOK_CMS_KEY}" \
  --data-urlencode 'tenant=big-health-company' \
  --data-urlencode 'sort=-publishedAt' \
  --data-urlencode 'limit=10'

Query	Description
`tenant`	Your tenant code. Or send `x-ovok-tenant-code`. A signed-in user may leave it out.
`projectId`	Your project's id, instead of the tenant code.
`environment`	`dev`, `staging` or `prod`. A wrong value is ignored and the host's default is used. See tenants and environments.
`locale`	`de`, `en`, `fr` or `es`. Leave it out for the default, English. See languages.
`page`	1 to 100. Default 1.
`limit`	1 to 50. Default 20.
`sort`	A top-level field name, `-` first for descending, such as `-publishedAt`. Leave it out for the CMS's own order.
`apiKey`	A CMS API key. Prefer the header; a key in a URL can end up in logs.

Response

{
  "docs": [
    {
      "id": 7,
      "slug": "privacy",
      "title": "Datenschutz",
      "body": { "root": { "type": "root", "children": [] } },
      "effectiveAt": "2026-09-01T00:00:00.000Z",
      "status": "published"
    }
  ],
  "totalDocs": 1,
  "limit": 20,
  "totalPages": 1,
  "page": 1,
  "pagingCounter": 1,
  "hasPrevPage": false,
  "hasNextPage": false,
  "prevPage": null,
  "nextPage": null
}

Field	Meaning
`docs`	The documents of this page. Only `published` ones.
`totalDocs`, `totalPages`	Counts across all pages, after the merge with the shared content.
`limit`, `page`, `pagingCounter`	The page you asked for.
`hasPrevPage`, `hasNextPage`, `prevPage`, `nextPage`	Paging links. `null` when there is none.

Each document has the fields of its collection:

Collection	Fields
`legal-pages`	`slug`, `title` and `body` (both localised; `body` is rich text), `effectiveAt`, `status`
`translations`	`slug` (the group), `title`, `strings` (a list of `key` and a localised `value`), `status`
`release-notes`	`slug`, `title`, `excerpt` and `body` (localised), `tags` (`announcement`, `new`, `improved`, `fixed`), `publishedAt`, `status`

Rich text is the editor's JSON structure, not HTML. Your own documents carry a tenant field; documents merged in from the shared tenant do not.

Order

Without sort, your documents come first, in the CMS's order, then the shared ones. With sort, the two lists are interleaved by that field: ties go to your document, and an empty value sorts last when ascending and first when descending. When a slug exists on both sides, only yours is listed.

What can go wrong

The errors in the overview apply. The checks run in this order, so the first failure is the one you see:

Authentication: 401, or 403 for another tenant.
The collection: 404 Not found for one that is not served.
locale, page, limit and sort: 400.
The tenant: 400.
The read: the CMS's own status, or 502 if it is unreachable.

Gotchas

Other query parameters are ignored. where[...], depth, draft and fields are not forwarded. Filtering is done in your app.
sort takes one top-level field. A dotted path, such as data.title, answers 400.
A page past the end is empty, not an error, with hasPrevPage true.
At most 5,000 documents are readable: 50 per page across 100 pages.
The merge stops at 200 documents. If your tenant or the shared one has more than 200 published documents in the collection, you get your own list, unmerged, with your own paging.
A failing shared read does not fail yours. You get your own documents with a 200.
The first read for a project can be slow. If the project has no content space yet, Ovok creates it, which can take a moment.
An edit takes up to a minute to show, and longer behind a CDN. See caching.

Who can call it​

Request​

Response​

Order​

What can go wrong​

Gotchas​

Related​