Ruby client for the Contentful Content Delivery API
Branch: allocation-hot…
Clone or download
Pull request Compare This branch is 10 commits ahead, 63 commits behind contentful:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.


Gem Version Build Status Coverage Status

Ruby client for the Contentful Content Delivery API.

Contentful is a content management platform for web applications, mobile apps and connected devices. It allows you to create, edit & manage content in the cloud and publish it anywhere via powerful API. Contentful offers tools for managing editorial teams and enabling cooperation between organizations.

IMPORTANT: We're collecting feedback before releasing version 2.0.0 of the SDK, if you're interested in helping, please drop by this issue and help us improving:


Add to your Gemfile and bundle:

gem 'contentful'


client =
  access_token: 'b4c0n73n7fu1',
  space: 'cfexampleapi'

If you plan on using the Preview API you need to specify the api_url:

client =
  access_token: 'b4c0n73n7fu1',
  space: 'cfexampleapi',
  api_url: ''

You can query for entries, assets, etc. very similar as described in the Delivery API Documentation. Please note, that all methods of the Ruby client library are snake_cased, instead of JavaScript's camelCase:

client.entry 'nyancat'

You can pass the usual filter options to the query:

client.entries(content_type: 'cat') # query for a content-type by its ID (not name)
client.entries('[ne]' => 'nyancat') # query for all entries except 'nyancat'
client.entries(include: 1) # include one level of linked resources
client.entries(content_type: 'cat', include: 1) # you can also combine multiple parameters

The results are returned as Contentful::Resource objects. Multiple results will be returned as Contentful::Array. The properties of a resource can be accessed through Ruby methods.

content_type = client.content_type 'cat'
content_type.description # "Meow."

System Properties behave the same and can be accessed via the #sys method. # => 'cat'
entry.type # => 'Entry'
asset.sys # { id: '...', type: '...' }

Entry Fields usually don't have direct method accessors, since they are based on individual content types. These fields can be accessed through the #fields method:

entry = client.entry 'nyancat'
entry.fields[:color] # rainbow

Please note, that no field type conversions will be done for entries by default.

Dynamic Entries

However, you can (and should) set :dynamic_entries to :auto in your client configuration. When using this option, the client will fetch all available content types and use them to create dynamic entries on the fly.

client =
  access_token: 'b4c0n73n7fu1',
  space: 'cfexampleapi',
  dynamic_entries: :auto

entry = client.entry 'nyancat' # => #<Contentful::DynamicEntry[cat]: ...>
entry.color # => 'rainbow'

Dynamic entries will have getter classes for the fields and do type conversions properly.

The :auto mode will fetch all content types on initialization. If you want to do it by hand later, you will need to set the option to :manual and call client.update_dynamic_entry_cache! to initialize all dynamic entries.

Using different locales

Entries can have multiple locales, by default, the client only fetches the entry with only its default locale. If you want to fetch a different locale you can do the following:

entries = client.entries(locale: 'de-DE')

Then all the fields will be fetched for the requested locale.

Contentful Delivery API also allows to fetch all locales, you can do so by doing:

entries = client.entries(content_type: 'cat', locale: '*')

# assuming the entry has a field called name
my_spanish_name = entries.first.fields('es-AR')[:name]

When requesting multiple locales, the object accessor shortcuts only work for the default locale.


Contentful::Array has an #each method that delegates to its items. It also includes Ruby's Enumerable module, providing methods like #min or #first. See the Ruby core documentation for further details.

Arrays also have a #next_page URL, which will rerun the request with a increased skip parameter, as described in the documentation.


You can easily request a resource that is represented by a link by calling #resolve:

happycat = client.entry 'happycat'
# => #<Contentful::Link: @sys={:type=>"Link", :linkType=>"Asset", :id=>"happycat"}>
happycat.fields[:image].resolve # => #<Contentful::Asset: @fields={ ...


There is a helpful method to add image resize options for an asset image:

# => "//
#     382a48dfa2cb16c47aa2c72f7b23bf09/happycatw.jpg"

client.asset('happycat').image_url width: 300, height: 200, format: 'jpg', quality: 100
# => "//
#     382a48dfa2cb16c47aa2c72f7b23bf09/happycatw.jpg?w=300&h=200&fm=jpg&q=100"

Resource Options

Resources, that have been requested directly (i.e. no child resources), can be fetched from the server again by calling #reload:

entries = client.entries
entries.reload # Fetches the array of entries again

Field Type "Object"

While for known field types, the field data is accessible using methods or the #fields hash with symbol keys, it behaves differently for nested data of the type "Object". The client will treat them as arbitrary hashes and will not parse the data inside, which also means, this data is indexed by Ruby strings, not symbols.

Client Configuration Options


Required option. The name of the space you want to access.


Required option. The space's secret token.


Whether to use https. Defaults to true.


How to authenticate with the API. Supports :header (default) or :query_string.


If set to true (default), error responses will be raised. If set to false, the error objects will simply be returned.


:auto or :manual. See resource description above for details on usage.


Defaults to false. If enabled, the API responses will not be parsed to resource objects. Might be useful for debugging.


See next paragraph for explanation


Enables gzip response content encoding, default to: true


Logging is disabled by default, it can be enabled by setting a logger instance and a logging severity.

client =
  access_token: 'b4c0n73n7fu1',
  space: 'cfexampleapi',
  logger: logger_instance,
  log_level: Logger::DEBUG

Example loggers:



The default severity is set to INFO and logs only the request attributes (headers, parameters and url). Setting it to DEBUG will also log the raw JSON response.


To be able to perform a request behind a proxy, you need to specify a :proxy_host. This can be a domain or IP address of the proxy server.


Specify the port number that is used by the proxy server for client connections.

:port_password, :port_username

To use the proxy with authentication, you need to specify port_username and port_password.


To increase or decrease the retry attempts after a 429 Rate Limit error. Default value is 1. Using 0 will disable retry behaviour. Each retry will be attempted after the value (in seconds) of the X-Contentful-RateLimit-Reset header, which contains the amount of seconds until the next non rate limited request is available, has passed. This is blocking per execution thread.


Maximum time to wait for next available request (in seconds). Default value is 60 seconds. Keep in mind that if you hit the houly rate limit maximum, you can have up to 60 minutes of blocked requests. It is set to a default of 60 seconds in order to avoid blocking processes for too long, as rate limit retry behaviour is blocking per execution thread.


Maximum amount of levels to resolve includes for SDK entities (this is independent of API-level includes - it represents the maximum depth the include resolution tree is allowed to resolved before falling back to Link objects). This include resolution strategy is in place in order to avoid having infinite circular recursion on resources with circular dependencies. Defaults to 20. Note: If you're using something like Rails::cache it's advisable to considerably lower this value (around 5 has proven to be a good compromise - but keep it higher or equal than your maximum API-level include parameter if you need the entire tree resolution).

Proxy example

client =
  access_token: 'b4c0n73n7fu1',
  space: 'cfexampleapi',
  proxy_host: '',
  proxy_port: 8080,
  proxy_username: 'username',
  proxy_password: 'secret_password',

Advanced Usage

Custom Resource Classes

You can define your own classes that will be returned instead of the predefined ones. Consider, you want to build a better Asset class. One way to do this is:

class MyBetterAsset < Contentful::Asset
  def https_image_url
    image_url.sub %r<\A//>, 'https://'

You can register your custom class on client initialization:

client =
  space: 'cfexampleapi',
  access_token: 'b4c0n73n7fu1',
  resource_mapping: {
    'Asset' => MyBetterAsset

More information on :resource_mapping can be found in examples/resource_mapping.rb and more on custom classes in examples/custom_classes.rb

You can also register custom entry classes to be used based on the entry's content_type using the :entry_mapping configuration:

class Cat < Contentful::Entry
  # define methods based on :fields, etc

client =
  space: 'cfexampleapi',
  access_token: 'b4c0n73n7fu1',
  entry_mapping: {
    'cat' => Cat

client.entry('nyancat') # is instance of Cat


The client also includes a wrapper for the synchronization endpoint. You can initialize it with the options described in the Delivery API Documentation or an URL you received from a previous sync:

client =
  access_token: 'b4c0n73n7fu1',
  space: 'cfexampleapi',
  default_locale: 'en-US'

sync = client.sync(initial: true, type: 'Deletion') # Only returns deleted entries and assets
sync = client.sync("") # Continues a sync

You can access the results either wrapped in Contentful::SyncPage objects:

sync.each_page do |page|
  # Find resources at: page.items

# More explicit version:
page = sync.first_page
until sync.completed?
  page = sync.next_page

Or directly iterative over all resources:

sync.each_item do |resource|
  # ...

When a sync is completed, the next sync url can be read from the Sync or SyncPage object:


Please note that synchronization entries come in all locales, so make sure, you supply a :default_locale property to the client configuration, when using the sync feature. This locale will be returned by default, when you call Entry#fields. The other localized data will also be saved and can be accessed by calling the fields method with a locale parameter:

first_entry = client.sync(initial: true, type: 'Entry').first_page.items.first
first_entry.fields('de-DE') # Returns German localizations


  • When an entry has related entries that are unpublished, they still end up in the resource as unresolved links. We consider this correct, because it is in line with the API responses and our other SDKs. However, you can use the workaround from issue #60 if you happen to want this working differently.

Migrating to 2.x

If you're a 0.x or a 1.x user of this gem, and are planning to migrate to the current 2.x branch. There are a few breaking changes you have to take into account:

  • Contentful::Link#resolve and Contentful::Array#next_page now require a Contentful::Client instance as a parameter.
  • Contentful::CustomResource does no longer exist, custom entry classes can now inherit from Contentful::Entry and have proper marshalling working.
  • Contentful::Resource does no longer exist, all resource classes now inherit from Contentful::BaseResource. Contentful::Entry and Contentful::Asset inherit from Contentful::FieldsResource which is a subclass of Contentful::BaseResource.
  • Contentful::DynamicEntry does no longer exist, if code checked against that base class, it should now check against Contentful::Entry instead.
  • Contentful::Client#dynamic_entry_cache (private) has been extracted to it's own class, and can be now manually cleared by using Contentful::ContentTypeCache::clear.
  • Contentful::BaseResource#sys and Contentful::FieldsResource#fields internal representation for keys are now snake cased to match the instance accessors. E.g. entry.fields[:myField] previously had the accessor entry.my_field, now it is entry.fields[:my_field]. The value in both cases would correspond to the same field, only change is to unify the style. If code accessed the values through the #sys or #fields methods, keys now need to be snake cased.
  • Circular references are handled as individual objects to simplify marshalling and reduce stack errors, this introduces a performance hit on extremely interconnected content. Therefore, to limit the impact of circular references, an additional configuration flag max_include_resolution_depth has been added. It is set to 20 by default (which corresponds to the maximum include level value * 2). This allows for non-circular but highly connected content to resolve properly. In very interconnected content, it also allows to reduce this number to improve performance. For a more in depth look into this you can read this issue.
  • #inspect now offers a clearer and better output for all resources. If your code had assertions based on the string representation of the resources, update to the new format <Contentful::#{RESOURCE_CLASS}#{additional_info} id="#{RESOURCE_ID}">.

For more information on the internal changes present in the 2.x release, please read the CHANGELOG


Copyright (c) 2014 Contentful GmbH - Jan Lelis. Copyright (c) 2016 Contentfuk GmbH - David Litvak.

See LICENSE.txt for further details.