Content Delivery API


The Content Delivery API (CDA), available at, is a read-only API for delivering content from Contentful to apps, websites and other media. Content is delivered as JSON data, and images, videos and other media as files.

The API is available via a globally distributed content delivery network (CDN). The server closest to the user serves all content, which minimizes latency and especially benefits mobile apps. Hosting content in multiple global data centers also improves the availability of content.


Any client requesting content from the CDA needs to provide an access token. You have two options to supply the access token, either as an Authorization request header field, or as an access_token URI query parameter. The CDA implements the standardized OAuth 2.0 bearer token specification already supported by many HTTP clients.

You create access tokens in the APIs tab of each space in the Contentful web app. Our reference guide has more details on how authentication works with Contentful.

API rate limits

Rate limits refer to the number of requests per unit of time you can make to Contentful APIs.

There are no limits enforced on requests that hit our CDN cache, i.e. the request doesn't count towards your rate limit and you can make an unlimited amount of cache hits. For requests that do hit the API, there is a rate limit of 216000 requests per hour and 78 requests per second. Higher rate limits might apply for enterprise clients.

The limit exists to prevent unlimited usage and restrict the consumption of resources from a single account. The rate limit uses a standard mechanism of the HTTP protocol to return a 429 response code with a X-Contentful-RateLimit-Reset header that tells a user when they can make their next request.

Header Status Code Description
X-Contentful-RateLimit-Reset 429. Returned only for rate limited requests. The number of seconds until the next request can be made.

Note: The X-Contentful-RateLimit-Reset header does not return the time until the next full reset. It tells you when you can make the next request. If a per second rate limit is enforced it will return 1 (time until the next second). If an hourly rate limit is enforced it will return the time until the next sliding window.

Common resource attributes

Every resource returned by the Content Delivery API will have a sys property, which is an object containing system managed metadata. The exact metadata available depends on the resource type, but at minimum it defines the sys.type property.

Note: None of the sys fields are editable and you can only specify the in the creation of an item (If it's not a *space_).

Contentful defines the property for every resource that is not a collection. For example, a Space resource will have a sys.type and

  "sys": {
    "type": "Space",
    "id": "yadj1kx9rmg0"
Field Type Description Applies to
sys.type String Resource type. All String Unique ID of resource. All except arrays Link Link to resource's space. Entries, assets, content types
sys.contentType Link Link to entry's content type. Entries
sys.revision Integer Published version of resource. Entries, assets, content types
sys.createdAt Date Time resource was created. Entries, assets, content types
sys.updatedAt Date Time resource was updated. Entries, assets, content types
sys.locale String Locale of the resource. Entries and assets

Note: The revision field refers to the current number of published revisions of an entry. Find out more in the Content Management API documentation.

Collection resources and pagination

Contentful returns collections of resources in a wrapper object that contains extra information useful for paginating over large result sets:

  "sys": { "type": "Array" },
  "skip": 0,
  "limit": 100,
  "total": 1256,
  "items": [ /* 100 individual resources */ ]

In the above example, a client retrieves the next 100 resources by repeating the same request, changing the skip query parameter to 100. You can use the order parameter, for example, order=sys.createdAt, when paging through larger result sets to keep ordering predictable.



All content and assets in Contentful belong to a space. You will generally have at least one space for a project, but use separate spaces for testing or staging.


Each space has a name, a set of locales, and metadata about the space.

Get a space

Content types

Defining a content type is a fundamental step in powering your applications with Contentful. A content type consists of a set of fields and other information, read this guide to learn more about modelling your content.

Content model

Get the content model of a space

Content type

Get a single content type


Entries represent anything defined as a Content Type in a space.

Entries collection

Get all entries of a Space


Get a single entry


Assets are the binary files in a space. An asset can be any file type and are usually attached to entries with links.

You can localize assets by providing separate files for each locale. Assets which are not localized are available as a single file under the default locale.

When you query for entries which contain links to assets then all assets are included by default.

Asset properties:

Field Type Description
sys Sys Common system properties.
fields.title Text Title of the asset.
fields.description Text Description of the asset.
fields.file File File(s) of the asset.
fields.file.fileName Symbol Original filename of the file.
fields.file.contentType Symbol Content type of the file.
fields.file.url Symbol URL of the file.
fields.file.details Object Details of the file, depending on its MIME type.
fields.file.details.size Number Size (in bytes) of the file.

For image assets, the fields.file.url field will point to For other file types, it will point to

You can use query parameters to define the image size, cropping parameters and other options. Find out more in our Images API reference.

Assets collection

Get all assets of a space


Get a single asset

Links are a powerful way to model relationships between content entries and assets. You can learn more about links in our concept guide.

Retrieval of linked items

When you have related content (e.g. entries with links to image assets) it's possible to include both search results and related data in a single request. Using the include parameter, you can specify the number of levels to resolve.

The maximum number of levels is 10. The API will throw a BadRequestError for higher values or values other than an integer. The default number, if the parameter is not specified, is 1. To omit linked items, specify include as 0.

If the standard items array has not already retrieved the linked entries, they will be in the includes.Entry array. Linked assets are inside the includes.Asset array.

Note: The include parameter resolves links between entries and assets within a space. Links between content types within a space are not included in the response.

Query entries

If you want to retrieve all items linked to a specific entry, the query URL should filter entries on their specific content_type, linking_field (field to link items) and entry_id from the target entry.

Query entries

Search parameters

You can use a variety of query parameters to search and filter items in the reponse from any collection endpoint including entries, content types and assets.

Content type

To search for entries with a specific content type, set the content_type URI query parameter to the ID you want to search for.

Note: When querying entries and using search filters based on fields or ordering by fields you must specify this parameter. You can only specify one value for content_type at a time.

Query entries

This example finds all entries of content type 'Product'.

Select operator

The select operator allows you to choose what fields to return from an entity. You can choose multiple values by combining comma separated operators.

For example, if you want to return the and of an Entry you would use:


You can fetch the entire sys or fields object and it's sub-fields by passing it to the select operator. For example to omit the sys object:


The select operator has some restrictions:

  • It is only applicable for collections of Entries and Assets, and with an Entry you must provide the content_type query parameter.

  • It can only select properties to a depth of 2. For example, select=fields.productName.en-US is not valid.

  • If you want to select a property for a specific locale, you need to combine the select and locale operators, e.g /assets/?select=fields.productName&locale=en-US.

  • You can select up to 100 properties.

  • If you use the select operator on an Entry with linked fields, only the content linked to a field you select will be returned.

If you provide an invalid property path, e.g fields.doesNotExist, Contentful returns a 400 Bad request containing the invalid property path.

Query entries

To select only the productName field of each entry.

Equality operator

You can search for exact matches by using the equality operator. This includes querying an entry by an ID value instead of retrieving the Entry directly, which allows you to include resolved links.

Note: Equality and inequality operators are not supported for text fields and you need to constrain search queries for fields with a content_type URI query parameter.

Query entries

To find all entries with the ID 5KsDBWseXY6QegucYAoacS (IDs are unique).

Inequality operator

Uses the [ne] parameter to exclude items matching a certain pattern.

Query entries

To return all entries, except those with the ID 5KsDBWseXY6QegucYAoacS.

Array equality/inequality

The equality/inequality operators also work with array fields:

  • Equality: If one of the items in an array matches the searched term, then it returns the entry.

  • Inequality: If one of the fields in an array matches the searched term, then the entry is not returned.

Note: As the query filters by a field, you need to specify a content type.

Query entries

This example finds all products tagged as accessories by matching fields.tags (an array) against a single value.

Array with multiple values

It's possible to use the [all] operator to retrieve entries matching a specific set of values (e.g. fields.likes[all]=flowers,accessories)

Query entries

To find all products tagged as flowers and accessories by using the all operator with fields.tags (an array) and the two values.


You can filter a field by multiple values with the [in] operator. When applied to an array value there must be at least one matching item. Similarly, when including a field value, you need to specify a Ccntent type.

Query entries

To retrieve entries that match accessories and flowers.


You can filter a field by multiple values with the [nin] operator. When applied to an array value there must be at least one not matching item. Similarly, when including a field value, you need to specify a content type.

Query entries

To retrieve all products tagged as neither 'flowers' nor 'accessories'.


You can check for the presence of a field using the [exists] operator. It checks if a certain field is defined (i.e. it has any value).

If the field is not assigned a value, it will not have a JSON payload and the operator will assume it to be 'non-existent'.

Query entries

To retrieve all entries that have a value for field.tags defined.


Four range operators are available that you can apply to date and number fields:

  • [lt]: Less than.

  • [lte]: Less than or equal to.

  • [gt]: Greater than.

  • [gte]: Greater than or equal to.

When applied to field values, you must specify the content type in the query.

Query entries

To retrieve entries updated since midnight of January 1st 2013.

It's possible to perform a full-text search across all text and symbol fields with the query parameter.

Note: Full-text search is case insensitive and might return more results than expected. A query will only take values with more than 1 character.

Query entries

To retrieve all entries containing the word 'design'.

Full-text search on a field

You can perform a full-text search on a specific field with the [match] operator.

Note: Full-text search is case insensitive and might return more results than expected. A query will only take values with more than 1 character.

Query entries

To retrieve all entries with '.com' in the 'website' field.

You can use a proximity search on location-enabled content to find results in a specified geographical area.

Note: Queries that include exact coordinates can't take advantage of our caching layer. With many use cases it should be enough to round the coordinates to 3 decimal places (an accuracy of about 300m), 2 decimal places (an accuracy of about 1km) or more to improve your cache hit rates.

Query entries

A common use case for location search is to search for places close to the user's current position.

Use the [near] operator to show results closest to a specific map location and order the results by distance.

This will return all entries sorted by distance from the point at latitude=38 and longitude=-122.

Locations in a bounding object

When displaying content on a map it's more resource efficient to retrieve only content that is in the current visible map area. For these cases, use the within operator.

Similar to the 'near me' use case, this lets you search for locations that are within the specific area on the map and can be useful for finding related entries in the vicinity of another entry.

There are two ways to search for nearby locations.

1. Using a bounding rectangle:

To search for locations within a rectangle area object, use the structure[within]=latitude1,longitude1,latitude2,longitude2, where:

  • latitude1 and longitude1 refer to the coordinates of the bottom left corner of the rectangle.

  • latitude2 and longitude2 refer to the coordinates of the top right corner of the rectangle.

2. Using a bounding circle:

The structure[within]=latitute,longitude,radius will return entries included in the circle with of the given latitude, longitude and radius (in km).

Query entries

To retrieve entries where is within the rectangle with:

  • Bottom left corner: latitude 1, longitude 2.

  • Top right corner: latitude 3, longitude 4.


You can order items in the response by specifying the order search parameter. You can use sys properties (such as sys.createdAt) or field values (such as fields.myCustomDateField) for ordering.

Note: You must set the content_type URI query parameter to the ID of the content type you want to filter by. You can only use the following field types with the order parameter:

Name JSON Primitive Description
Symbol String Basic list of characters. Maximum length is 256.
Integer Number Number type without decimals. Values from -253 to 253.
Number Number Number type with decimals.
Date String Date/time in ISO-8601 format.
Boolean Boolean Flag, true or false.

The following field types do not support the order parameter:

Name JSON Primitive Description
Text String Same as symbol, but can be filtered via full-text search. Maximum length is 50,000.
Link Object See links
Array Array List of values. Value type depends on field.items.type.
Object Object Arbitrary object.

Query entries

To retrieve all entries ordered by creation date.

Reverse order

You can reverse the sort order by prefixing the field with a - symbol.

Query entries

To order results by the time of the last update, newest to oldest.

Order with multiple parameters

You can order items by specifying the order parameter with attributes(attribute, attribute2). Prefix the field with a - sign to reverse the sort order of the attribute.

Query entries


You can specify the maximum number of results as a limit search parameter.

Note: The maximum number of entries returned by the API is 1000. The API will throw a BadRequestError for values higher than 1000 and values other than an integer. The default number of entries returned by the API is 100.

Query entries

To limit results to 3 entries.


You can specify an offset with the skip search parameter.

Note: The API will throw a BadRequestError for values less than 0 or values other than an integer.

By combining skip and limit you can paginate through results:

Page 1: skip=0, limit=15 Page 2: skip=15, limit=15 Page 3: skip=30, limit=15 etc.

Query entries

To skips 3 entries.

Filtering assets by MIME type

You can filter assets by their MIME type group by using the mimetype_group query parameter. Valid groups are attachment, plaintext, image, audio, video, richtext, presentation, spreadsheet, pdfdocument, archive, code and markup. By default, the API will return all assets.

Query assets

To return only image assets.

Search on references

You can search for entries based on the values of referenced entries.

For example, if you want to find products of a particular brand, you can use search on references to find the products with one API request.

Structure of a query

Here's how the example above would look as a query:

  • First is the content_type parameter which you must include when you want to filter based on the values of the fields of entries of a certain content type.

  • Second is which you use to to filter on fields of entries from content type 2PqfXUJwE8qSYKuM0U6w8M.

  • Third is fields.brand.fields.companyName[match]=Lemnos which filters entries to those branded as 'Lemnos'.

Note: All filters on referenced entries are scoped with the path to the field that contains the reference. In the previous example that path is fields.brand.

You can search on multiple references at once and combine them into one singe API query with a series of field and value pairs.

Available nested filters

You can use the all, in, nin, exists, match, gt, gte,lt, lte, ne, near and within filters when searching on references. The semantics of these filters are the same as when you use them on a non referenced entry.


Searching on references has the following restrictions:

  • You can only search on references on fields which link to a single entry. Fields which hold references to many entries or fields with references to assets are not supported.

  • The maximum number of reference searches in one query is 2. A larger value will return an error.

Search on references

This example will return the Lemnos branded products.


Retrieve localized entries

You can specify a locale for entries using the locale URI query parameter.

The locale parameter must be the code of a locale in the space you're querying, or the wildcard value *. If you don't specify a locale, the default locale of the space is used.

If there's no content available for the requested locale the API will try the fallback locale of the requested locale.

For example you have the de-CH (Swiss German) locale and configure it to fallback to de-DE (German). When you request content for de-CH any missing field in that locale will be replace with its fallback value in de-DE (if they exist). For more information about locale fallbacks read the locale section of the CMA docs.

When you specify locale=*, field values are nested in an object with keys corresponding to each locale with a defined value. Instead of writing fields.productName when accessing the response data from your code, use fields.productName[localeCode].

If the result contains only a single locale, resources will include the property sys.locale indicating the locale of that object.

Learn more about locales in our concepts document.

Query entries

fields.productName is the only localized field in the product content type, so returns all other fields in the default locale.


The Sync API allows you to keep a local copy of all content in a space up-to-date via delta updates, or only the content that has changed.

Syncing specific content

By default a sync request response includes all resource types including assets, entries of all content types and deleted resources.

To sync specific content you can specify a type parameter for the initial sync:

Parameter Description
all (default) Include all new and changed content, i.e., assets, entries and deletions.
Asset Only include new and changed assets.
Entry Only include new and changed entries.
Deletion Only include deletions of assets and entries.
DeletedAsset Only include deletions of assets.
DeletedEntry Only include deletions of entries.

Initial synchronization

A client syncs content by using the sync endpoint. A single sync consists of one or more pages, each a separate request and response.

The first time a client synchronizes content it sends a request to the sync endpoint with the URI query parameter initial set to true.

You should only sync with initial=true for the initial sync when a client has no existing content. This transfers all content from a space, but does not contain content deleted from fields or deleted assets and content entries.

Query entries

Initial synchronization of entries of a specific content type

For entries, you can also specify a content_type parameter. When specifying content_type you must specify type as Entry (meaning there will be no deletions). You can only specify the type and content_type parameter at the initial sync along with the initial parameter. Any subsequent syncs will only include the types you have specified. If you want to sync entries by content type you should separately subscribe to a Deletion or DeletedEntry sync to be notified when entries are deleted.

Query entries

Pagination and subsequent syncs

Synchronization responses contain either a nextPageUrl or nextSyncUrl attribute. If there is more content for the current sync, nextPageUrl will contain an URL with a sync_token which will retrieve the next page. This continues until you have retrieved all content.

At this point, the response will contain a nextSyncUrl which you use to make requests in the future and retrieve delta updates between the current content on Contentful and what you retrieved with your last sync request. Upon completion, you will receive a new nextSyncUrl which you can again use for future updates.

If a client loses the nextSyncUrl, the client should delete its local content and perform an initial synchronization again.

Query entries