A long-lived project that still receives updates
Ruby wrapper for the REST API at https://www.zendesk.com. Documentation at https://developer.zendesk.com.


> 2.0.0
>= 3.5.2, < 6.0.0
 Project Readme

Zendesk API Client

Test Gem Version Code Climate


This Ruby gem is a generic wrapper around Zendesk's REST API. Follow this README and the wiki for how to use it.

You can interact with all the resources defined in resources.rb. Basically we have some cleaver code to convert Ruby objects into HTTP requests.

Please refer to our API documentation for the specific endpoints and once you understand the mapping between Ruby and the HTTP endpoints you should be able to call any endpoint.

The Yard generated documentation is available in at RubyDoc.

Please report any bug in the Github issues page.

You might want to try out this gem in a REPL for exploring your options, if so, check out this project.

Product Support

This Ruby gem supports the REST API's for Zendesk Support, Zendesk Guide, and Zendesk Talk. It does not yet support other Zendesk products such as Zendesk Chat, Zendesk Explore, and Zendesk Sell.


The Zendesk API client can be installed using Rubygems or Bundler.


gem install zendesk_api


Add it to your Gemfile

gem "zendesk_api"

Then bundle as usual.


Configuration is done through a block returning an instance of ZendeskAPI::Client.

require 'zendesk_api'

client = ZendeskAPI::Client.new do |config|
  # Mandatory:

  config.url = "<- your-zendesk-url ->" # e.g. https://yoursubdomain.zendesk.com/api/v2

  # Basic / Token Authentication
  config.username = "login.email@zendesk.com"

  # Choose one of the following depending on your authentication choice
  # More information on obtaining API tokens can be found here:
  # https://developer.zendesk.com/api-reference/introduction/security-and-auth/#api-token
  config.token = "your zendesk token"

  # OAuth Authentication
  # More information on obtaining OAuth access tokens can be found here:
  # https://developer.zendesk.com/api-reference/introduction/security-and-auth/#oauth-access-token
  config.access_token = "your OAuth access token"

  # Optional:

  # Retry uses middleware to notify the user
  # when hitting the rate limit, sleep automatically,
  # then retry the request.
  config.retry = true

  # Raise error when hitting the rate limit.
  # This is ignored and always set to false when `retry` is enabled.
  # Disabled by default.
  config.raise_error_when_rate_limited = false

  # Logger prints to STDERR by default, to e.g. print to stdout:
  require 'logger'
  config.logger = Logger.new(STDOUT)

  # Disable resource cache (this is enabled by default)
  config.use_resource_cache = false

  # Changes Faraday adapter
  # config.adapter = :patron

  # Merged with the default client options hash
  # config.client_options = {:ssl => {:verify => false}, :request => {:timeout => 30}}

  # When getting the error 'hostname does not match the server certificate'
  # use the API at https://yoursubdomain.zendesk.com/api/v2

  # Change retry configuration (this is disabled by default)
  config.retry_on_exception = true

  # Error codes when the request will be automatically retried. Defaults to 429, 503
  config.retry_codes = [ 429 ]


The result of configuration is an instance of ZendeskAPI::Client which can then be used in two different methods.

One way to use the client is to pass it in as an argument to individual classes.

Note: all method calls ending in ! will raise an exception when an error occurs, see the wiki page for more info.

ZendeskAPI::Ticket.new(client, :id => 1, :priority => "urgent") # doesn't actually send a request, must explicitly call #save!

ZendeskAPI::Ticket.create!(client, :subject => "Test Ticket", :comment => { :value => "This is a test" }, :submitter_id => client.current_user.id, :priority => "urgent")
ZendeskAPI::Ticket.find!(client, :id => 1)
ZendeskAPI::Ticket.destroy!(client, :id => 1)

You can also update ticket objects.

ticket = ZendeskAPI::Ticket.find!(client, :id => 1)
ticket.update(:comment => { :value => "This is a test reply." })


Another way is to use the instance methods under client.

client.tickets.find!(:id => 1)
client.tickets.build(:subject => "Test Ticket")
client.tickets.create!(:subject => "Test Ticket", :comment => { :value => "This is a test" }, :submitter_id => client.current_user.id, :priority => "urgent")
client.tickets.destroy!(:id => 1)

The methods under ZendeskAPI::Client (such as .tickets) return an instance of ZendeskAPI::Collection, a lazy-loaded list of that resource. Actual requests may not be sent until an explicit ZendeskAPI::Collection#fetch!, ZendeskAPI::Collection#to_a!, or an applicable methods such as #each.


Resource updating is implemented by sending only the changed? attributes to the server (see ZendeskAPI::TrackChanges). Unfortunately, this module only hooks into Hash meaning any changes to an Array not resulting in a new instance will not be tracked and sent.

zendesk_api_client_rb $ bundle console
> a = ZendeskAPI::Trackie.new(:test => []).tap(&:clear_changes)
> a.changed?(:test)
 => false
> a.test << "hello"
 => ["hello"]
> a.changed?(:test)
 => false
> a.test += %w{hi}
 => ["hello", "hi"]
> a.changed?(:test)
 => true


ZendeskAPI::Collections can be paginated:

# Note that CBP (cursor based pagination) is the default and preferred way
# and has fewer limitations on deep pagination
tickets = client.tickets.per_page(3)
page1 = tickets.fetch! # GET /api/v2/tickets?page[after]={cursor}&page[size]=3
page2 = tickets.next # GET /api/v2/tickets?page[after]={cursor}&page[size]=3
# ...

# OR...
# Note that OBP (offset based pagination) can incur to various limitations
tickets = client.tickets.page(2).per_page(3)

next_page = tickets.next # => 3
tickets.fetch! # GET /api/v2/tickets?page=3&per_page=3
previous_page = tickets.prev # => 2
tickets.fetch! # GET /api/v2/tickets?page=2&per_page=3

Iteration over all resources and pages is handled by Collection#all:

client.tickets.all! do |resource|
  # every resource, from all pages, will be yielded to this block

If given a block with two arguments, the page number is also passed in.

client.tickets.all! do |resource, page_number|
  # all resources will be yielded along with the page number

Cursor Based Pagination

A few endpoints related to organizations, tickets, triggers and groups will now make use of cursor based pagination by default. It is also recommended to use CBP whenever the Zendesk developer documentation says it's supported. Pass page[size]=number in the parameters to attempt a CBP request, like the example below:

client.connection.get('/api/v2/suspended_tickets', page: { size: 100 }).body
{"suspended_tickets"=>[...], "meta"=>{"has_more"=>true, "after_cursor"=>" ... ", "before_cursor"=>nil}, "links"=>{"prev"=>nil, "next"=>"..."}}


Callbacks can be added to the ZendeskAPI::Client instance and will be called (with the response env) after all response middleware on a successful request.

client.insert_callback do |env|
  puts env[:response_headers]

Resource management

Individual resources can be created, modified, saved, and destroyed.

ticket = client.tickets[0] # ZendeskAPI::Ticket.find(client, :id => 1)
ticket.priority = "urgent"
ticket.attributes # => { "priority" => "urgent" }
ticket.save! # Will PUT => true
ticket.destroy! # => true

ZendeskAPI::Ticket.new(client, { priority: "urgent" })
ticket.new_record? # => true
ticket.save! # Will POST


To facilitate a smaller number of requests and easier manipulation of associated data we allow "side-loading," or inclusion, of selected resources.

For example: A ZendeskAPI::Ticket is associated with ZendeskAPI::User through the requester_id field. API requests for that ticket return a structure similar to this:

"ticket": {
  "id": 1,
  "url": "http.....",
  "requester_id": 7,

Calling ZendeskAPI::Ticket#requester automatically fetches and loads the user referenced above (/api/v2/users/7). Using side-loading, however, the user can be partially loaded in the same request as the ticket.

tickets = client.tickets.include(:users)
# Or client.tickets(:include => :users)
# Does *NOT* make a request to the server since it is already loaded
tickets.first.requester # => #<ZendeskAPI::User id=...>

# OR

ticket = client.tickets.find!(:id => 1, :include => :users)
ticket.requester # => #<ZendeskAPI::User id=...>

Currently, this feature is limited to only a few resources and their associations. They are documented on developer.zendesk.com.

Recommended Approach

For better control over your data and to avoid large response sizes, consider fetching related resources explicitly. This approach can help you manage data loading more precisely and can lead to optimized performance for complex applications.

ticket = ZendeskAPI::Ticket.find(id: 1)
requester = ZendeskAPI::User.find(id: ticket.requester_id)

By explicitly fetching associated resources, you can ensure that your application only processes the data it needs, improving overall efficiency.


Support for the Agent Availability API

An agent’s availability includes their state (such as online) for each channel (such as messaging), and their unified state across channels. It also includes the work items assigned to them.

# All agent availabilities

# fetch availability for one agent, their channels and work items
agent_availability = ZendeskAPI::AgentAvailability.find(client, 386390041152)

# Using the agent availability filter
ZendeskAPI::AgentAvailability.search(client, { select_channel: 'support' })
ZendeskAPI::AgentAvailability.search(client, { channel_status: 'support:online' })


Searching is done through the client. Returned is an instance of ZendeskAPI::Collection:

client.search(:query => "my search query") # /api/v2/search.json?query=...
client.users.search(:query => "my new query")  # /api/v2/users/search.json?query=...

Special case: Custom resources paths

API endpoints such as tickets/recent or topics/show_many can be accessed through chaining. They will too return an instance of ZendeskAPI::Collection.

client.topics.show_many(:verb => :post, :ids => [1, 2, 3])

Special Case: Current user

Use either of the following to obtain the current user instance:

client.users.find!(:id => 'me')

Special Case: Importing a ticket

Bulk importing tickets allows you to move large amounts of data into Zendesk.

ticket = ZendeskAPI::Ticket.import(client, :subject => "Help", :comments => [{ :author_id => 19, :value => "This is a comment" }])

Further documentation can be found on developer.zendesk.com

Attaching files

Files can be attached to ticket comments using either a path or the File class and will be automatically uploaded and attached.

ticket = ZendeskAPI::Ticket.new(client, :comment => { :value => "attachments" })
ticket.comment.uploads << "img.jpg"
ticket.comment.uploads << File.new("img.jpg")

Apps API

v1.1.0 introduces support for the Zendesk Apps API

Creating Apps

upload = client.apps.uploads.create!(:file => "path/to/app.zip")
client.apps.create!(:name => "test", :upload_id => upload.id)

# Or

app = ZendeskAPI::App.new(client, :name => "test")
app.upload = "path/to/app.zip"

# Or

upload = ZendeskAPI::App::Upload.new(client, :file => "path/to/app.zip")

app = ZendeskAPI::App.new(client, :name => "test")
app.upload_id = upload.id

# Or

client.apps.create!(:name => "test", :upload => "app.zip")

Note: job statuses are currently not supported, so you must manually poll the job status API for app creation.

body = {}
until %w{failed completed}.include?(body["status"])
  response = client.connection.get(app.response.headers["Location"])
  body = response.body


Updating Apps

upload = client.apps.uploads.create!(:file => "NewApp.zip")

# Then

client.apps.update!(:id => 123, :upload_id => upload.id)

# Or

app = ZendeskAPI::App.new(client, :id => 123)
app.upload_id = upload.id

# Or

ZendeskAPI::App.update!(client, :id => 123, :upload_id => upload.id)

Deleting Apps

client.apps.destroy!(:id => 123)

app = ZendeskAPI::App.new(client, :id => 123)

ZendeskAPI::App.destroy!(client, :id => 123)

Installing an App

Installation name is required

installation = ZendeskAPI::AppInstallation.new(client, :app_id => 123, :settings => { :name => 'Name' })

# or

client.apps.installations.create!(:app_id => 123, :settings => { :name => 'Name' })

# or

ZendeskAPI::AppInstallation.create!(client, :app_id => 123, :settings => { :name => 'Name' })

List Installations

apps = client.app.installations

Update Installation

client.app.installations.update!(:id => 123, :settings => { :title => "My New Name" })

installation = ZendeskAPI::AppInstallation.new(client, :id => 123)
installation.settings = { :title => "My New Name" }

ZendeskAPI::AppInstallation.update!(client, :id => 123, :settings => { :title => "My New Name" })

Delete Installation

client.app.installations.destroy!(:id => 123)

installation = ZendeskAPI::AppInstallation.new(client, :id => 123)

ZendeskAPI::AppInstallation.destroy!(client, :id => 123)

Running the gem locally

See .github/workflows/main.yml to understand the CI process.

bundle exec rake # Runs the tests
bundle exec rubocop # Runs the lint (use `--fix` for autocorrect)

Releasing a new gem version

  1. From updated master: git checkout -b bump-vX.X.X, according to SemVer
  2. Ensure the CHANGELOG is correct and updated, this is your last opportunity
  3. Execute bundle exec bump:patch # minor|major, this bumps the version in a new commit, and adds the relative git tag
  4. Push to GitHub git push origin vX.X.X -u && git push --tags
  5. Raise a PR (example) including the code diff (example)
  6. Get it approved and merged
  7. Post a message in Slack #rest-api (example TODO), so advocacy are aware that we are going to release a new gem, just in case any customer complains about something related to the gem
  8. After 2 hours from the above message, you can approve the release of the gem


  1. Fork the project.
  2. Make your feature addition or bug fix.
  3. Add tests for it. This is important so that we don't break it in a future version unintentionally.
  4. Commit. Do not alter Rakefile, version, or history. (If you want to have your own version, that is fine, but bump version in a commit by itself that we can ignore when we pull.)
  5. Submit a pull request.

Note: Live specs will likely fail for external contributors. The Zendesk devs can help with that. If you have permissions and some live specs unexpectedly fail, that might be a data error, see the REPL for that.

Merging contributors pull requests

External contributions don't run live specs, so we need to use a workaround. Assuming a PR from author:author/branch to default_branch:

  1. Create a branch in our repo author_branch
  2. Change the destination branch of the PR to author_branch
  3. Merge
  4. Create a pr in our repo from default_branch
    • Make sure they know the commits still carry their name
    • Example

Copyright and license

Copyright 2015-2023 Zendesk