A Ruby client for the Slack Web, RealTime Messaging and Events APIs. Comes with a handy command-line client, too. If you are not familiar with these concepts, you might want to watch this video.

• Useful to Me?
• Stable Release
• Installation
• Usage
  • Create a New Bot Integration
  • OAuth Code Grant
  • Using the Legacy API Token
  • Global Settings
  • Web Client
    • Web Client Examples
      • Test Auth
      • Send Messages
      • List Channels
      • Upload a File
      • Get Channel Info
      • Get User Info
      • Search for a User
      • Other
    • Web Client Options
    • Pagination Support
    • Error Handling
      • Slack Errors
      • Rate Limiting
      • Other Errors
  • RealTime Client
    • Slack::RealTime::Stores::Store
    • Slack::RealTime::Stores::Starter
  • Configuring Slack::RealTime::Client
    • Caveats
      • websocket_ping
  • Connection Methods
  • Combining RealTime and Web Clients
  • Large Team Considerations
    • Concurrency
      • Async
      • Faye::Websocket with Eventmachine
      • Celluloid
  • Events API
    • Configuring Slack::Events
    • Verifying the Request Signature
  • Message Parsing
  • Command-Line Client
    • Authenticate with Slack
    • Send a Message
    • Get Channel Id
    • Get Channel Info
    • List Users
• History
• Contributing
• Copyright and License

• This library will let you send messages to Slack via the Web API, send and receive messages via the Real Time Messaging API and facilitate integration with the Events API.
• To respond to slash commands, interactive components or events, implement a web application using your favorite web framework and use this library to call the Slack Web API and to verify that events are coming from Slack.
• To build a bot using the Real Time Messaging API, use slack-ruby-bot, which uses this library.
• To roll out a complete service using the Real Time Messaging API with Slack button integration to multiple teams, check out slack-ruby-bot-server, which is built on top of slack-ruby-bot, which uses this library.

You're reading the documentation for the next release of slack-ruby-client. Please see the documentation for the last stable release, v0.14.5 unless you're integrating with HEAD. See UPGRADING when upgrading from an older version.

Add to Gemfile.

gem 'slack-ruby-client'

If you're going to be using the RealTime client, add either async-websocket, eventmachine and faye-websocket or celluloid-io. See below for more information about concurrency. We recommend you use async-websocket.

gem 'async-websocket', '~> 0.8.0'

Run bundle install.

To integrate your bot with Slack, you must first create a new Slack App.

Once created, go to the app's Basic Info tab and grab the Client ID and Client Secret. You'll need these in order complete an OAuth code grant flow as described at slack-ruby-bot-server.

Although OAuth is recommended, you can still generate a legacy API token for your app and use it for some interactions.

Slack.configure do |config|
  config.token = ENV['SLACK_API_TOKEN']
end

This sets a global default token. You can also pass a token into the initializer of both Slack::Web::Client and Slack::RealTime::Client or configure those separately via Slack::Web::Config.configure and Slack::RealTime::Config.configure. The instance token will be used over the client type token over the global default.

The following global settings are supported via Slack.configure.

The Slack Web API allows you to build applications that interact with Slack.

Here are some examples of how to use the web client with the Web API.

client = Slack::Web::Client.new
client.auth_test

Send messages with chat_PostMessage.

client.chat_postMessage(channel: '#general', text: 'Hello World', as_user: true)

See a fully working example in examples/hi_web.

List channels with channels_list.

channels = client.channels_list.channels

general_channel = channels.detect { |c| c.name == 'general' }

Upload a file with files_upload.

client.files_upload(
  channels: '#general',
  as_user: true,
  file: Faraday::UploadIO.new('/path/to/avatar.jpg', 'image/jpeg'),
  title: 'My Avatar',
  filename: 'avatar.jpg',
  initial_comment: 'Attached a selfie.'
)

You can use a channel ID or name (prefixed with #) in all functions that take a :channel argument. Lookup by name is not supported by the Slack API and the channels_id method called invokes channels_list in order to locate the channel ID.

client.channels_info(channel: 'C04KB5X4D') # calls channels_info

client.channels_info(channel: '#general') # calls channels_list followed by channels_info

You can use a user ID or name (prefixed with @) in all functions that take a :user argument. Lookup by name is not supported by the Slack API and the users_id method called invokes users_list in order to locate the user ID.

client.users_info(user: 'U092BDCLV') # calls users_info

client.users_info(user: '@dblock') # calls users_list followed by users_info

Constructs an in-memory index of users and searches it. If you want to use this functionality, add the picky gem to your project's Gemfile.

client.users_search(user: 'dblock')

Refer to the Slack Web API Method Reference for the list of all available functions.

You can configure the Web client either globally or via the initializer.

Slack::Web::Client.configure do |config|
  config.user_agent = 'Slack Ruby Client/1.0'
end

client = Slack::Web::Client.new(user_agent: 'Slack Ruby Client/1.0')

The following settings are supported.

You can also pass request options, including timeout and open_timeout into individual calls.

client.channels_list(request: { timeout: 180 })

You can also control what proxy options are used by modifying the http_proxy environment variable per Net::HTTP's documentation.

Note that Docker on OSX seems to incorrectly set the proxy, causing Faraday::ConnectionFailed, ERROR -- : Failed to open TCP connection to : (getaddrinfo: Name or service not known). You might need to manually unset http_proxy in that case, eg. http_proxy="" bundle exec ruby ./my_bot.rb.

The Web client natively supports cursor pagination for methods that allow it, such as users_list. Supply a block and the client will make repeated requests adjusting the value of cursor with every response. The default limit is set to 100 and can be adjusted via Slack::Web::Client.config.default_page_size or by passing it directly into the API call.

all_members = []
client.users_list(presence: true, limit: 10) do |response|
  all_members.concat(response.members)
end
all_members # many thousands of team members retrieved 10 at a time

When using cursor pagination the client will automatically pause and then retry the request if it runs into Slack rate limiting. (It will pause according to the Retry-After header in the 429 response before retrying the request.) If it receives too many rate-limited responses in a row it will give up and raise an error. The default number of retries is 100 and can be adjusted via Slack::Web::Client.config.default_max_retries or by passing it directly into the method as max_retries.

You can also proactively avoid rate limiting by adding a pause between every paginated request with the sleep_interval parameter, which is given in seconds.

all_members = []
client.users_list(presence: true, limit: 10, sleep_interval: 5, max_retries: 20) do |response|
  # pauses for 5 seconds between each request
  # gives up after 20 consecutive rate-limited responses
  all_members.concat(response.members)
end
all_members # many thousands of team members retrieved 10 at a time

If Slack returns an error for the request, then an error will be raised. The error class is specific to the type of error that Slack returns. For instance if Slack returns account_inactive then the error will be Slack::Web::Api::Errors::AccountInactive. This allows you to handle certain types of errors as needed:

rescue Slack::Web::Api::Errors::AccountInactive => e
  # deal with inactive account
end

All of these errors inherit from Slack::Web::Api::Errors::SlackError, so you can handle or silence all errors if necessary:

rescue Slack::Web::Api::Errors::SlackError => e
  # capture all Slack errors
end

If there's a new error type that is not yet known by this library, then it will raise Slack::Web::Api::Errors::SlackError. (Update the Web API if you find that errors are missing — see CONTRIBUTING.)

In all of these cases the error message contains the error code, which is also accessible with slack_error.error. In case of multiple errors, the error message contains the error codes separated by commas, or they are accessible as an array with slack_error.errors. The original response is also accessible using the response attribute. The response_metadata is accessible with slack_error.response_metadata.

If you exceed Slack’s rate limits, a Slack::Web::Api::Errors::TooManyRequestsError will be raised instead. (This does not inherit from Slack::Web::Api::Errors::SlackError.)

In any other case, a Faraday::ClientError will be raised. This may be the case if Slack is temporarily unavailable, for example.

The Real Time Messaging API is a WebSocket-based API that allows you to receive events from Slack in real time and send messages as user.

client = Slack::RealTime::Client.new

client.on :hello do
  puts "Successfully connected, welcome '#{client.self.name}' to the '#{client.team.name}' team at https://#{client.team.domain}.slack.com."
end

client.on :message do |data|
  case data.text
  when 'bot hi' then
    client.message(channel: data.channel, text: "Hi <@#{data.user}>!")
  when /^bot/ then
    client.message(channel: data.channel, text: "Sorry <@#{data.user}>, what?")
  end
end

client.on :close do |_data|
  puts "Client is about to disconnect"
end

client.on :closed do |_data|
  puts "Client has disconnected successfully!"
end

client.start!

You can send typing indicators with typing.

client.typing channel: data.channel

You can send a ping with ping.

client.ping

By default, the RealTime client exposes and maintains a local store with the properties of rtm.start upon a successful connection.

It also tracks changes, such as users being renamed, added or deleted, therefore client.users is always up-to-date.

Tracking with a local store can be disabled with Slack::RealTime::Client.new(store_class: nil). Other stores are also available.

The default store that tracks all changes. By default the client will be connected using rtm_start.

A smaller store that only stores and tracks information about the bot user, but not channels, users, groups, ims or bots. By default the client will be connected using rtm_connect.

You can configure the RealTime client either globally or via the initializer.

Slack::RealTime::Client.configure do |config|
  config.websocket_ping = 42
end

client = Slack::RealTime::Client.new(websocket_ping: 42)

The following settings are supported.

Note that the RealTime client uses a Web client to obtain the WebSocket URL via rtm.start or rtm.connect. While token and logger options are passed down from the RealTime client, you may also configure Web client options via Slack::Web::Client.configure as described above.

See a fully working example in examples/hi_real_time.

This setting determines how long the socket can be idle before sending a ping message to confirm it's still connected.

It's important to note that if a ping message was sent and no response was received within the amount of time specified in websocket_ping the client will attempt to reestablish it's connection to the message server.

Note that the ping may take between websocket_ping and websocket_ping * 3/2 seconds to actually trigger when there is no activity on the socket. This is because the timer that checks whether to ping is triggered at every websocket_ping / 2 interval.

To disable this feature set websocket_ping to 0.

The RealTime client uses either rtm.start or rtm.connect to open a connection. The former retrieves a lot of team information while the latter only serves connection purposes and is preferred. You should let the library choose the right method for you based on the store_class used and override this behavior with start_method when necessary.

Slack::RealTime::Client.configure do |config|
  config.start_method = :rtm_start
end

Since the Web client is used to obtain the RealTime client's WebSocket URL, you can continue using the Web client in combination with the RealTime client.

client = Slack::RealTime::Client.new

client.on :message do |data|
  case data.text
  when 'bot hi' then
    client.web_client.chat_postMessage(channel: data.channel, text: "Hi <@#{data.user}>!")
  when /^bot/ then
    client.web_client.chat_postMessage(channel: data.channel, text: "Sorry <@#{data.user}>, what?")
  end
end

client.start!

See a fully working example in examples/hi_real_time_and_web.

The rtm.start call downloads a large amount of data. For large teams, consider reducing the amount of unnecessary data downloaded with start_options. You may also want to increase the default timeout of 180 seconds.

Slack::RealTime::Client.configure do |config|
  # Return timestamp only for latest message object of each channel.
  config.start_options[:simple_latest] = true
  # Skip unread counts for each channel.
  config.start_options[:no_unreads] = true
  # Increase request timeout to 6 minutes.
  config.start_options[:request][:timeout] = 360
end

See #134 for a discussion on this topic.

Slack::RealTime::Client needs help from a concurrency library and supports Async, Faye::WebSocket with Eventmachine and Celluloid. It will auto-detect one or the other depending on the gems in your Gemfile, but you can also set concurrency explicitly.

Slack::RealTime.configure do |config|
  config.concurrency = Slack::RealTime::Concurrency::Async
end

Use client.start_async instead of client.start!. A good example of such application is slack-ruby-bot-server.

client = Slack::RealTime::Client.new

client.start_async

This is the recommended library. Add async-websocket to your Gemfile.

gem 'async-websocket'

See a fully working example in examples/hi_real_time_async_async.

Add the following to your Gemfile.

gem 'faye-websocket'

See a fully working example in examples/hi_real_time_async_eventmachine.

Add the following to your Gemfile.

gem 'celluloid-io', require: ['celluloid/current', 'celluloid/io']

See a fully working example in examples/hi_real_time_async_celluloid.

This library provides limited support for the Slack Events API.

You can configure Events support globally.

Slack::Events.configure do |config|
  config.signing_secret = 'secret'
end

The following settings are supported.

Slack signs its requests using a secret that's unique to your app. Verify incoming HTTP requests as follows.

slack_request = Slack::Events::Request.new(http_request)
slack_request.verify!

To specify secrets on a per-request basis:

Slack::Events::Request.new(http_request,
                           signing_secret: signing_secret,
                           signature_expires_in: signature_expires_in)

The verify! call may raise Slack::Events::Request::MissingSigningSecret, Slack::Events::Request::InvalidSignature or Slack::Events::Request::TimestampExpired errors.

All text in Slack uses the same system of escaping: chat messages, direct messages, file comments, etc. Use Slack::Messages::Formatting to unescape incoming messages. This comes handy, for example, you want to treat all input to a real time bot as plain text.

Slack::Messages::Formatting.unescape('Hello & <world>'))
  # => 'Hello & <world>'
Slack::Messages::Formatting.unescape('Hey <@U024BE7LH|bob>, did you see my file?'))
  # => 'Hey @bob, did you see my file?'
Slack::Messages::Formatting.unescape('Hey <@U02BEFY4U>'))
  # => 'Hey @U02BEFY4U'
Slack::Messages::Formatting.unescape('This message contains a URL <http://foo.com/>'))
  # => 'This message contains a URL http://foo.com/'
Slack::Messages::Formatting.unescape('So does this one: <http://www.foo.com|www.foo.com>'))
  # => 'So does this one: www.foo.com'
Slack::Messages::Formatting.unescape('<mailto:[email protected]|Bob>'))
  # => 'Bob'
Slack::Messages::Formatting.unescape('Hello <@U123|bob>, say hi to <!everyone> in <#C1234|general>'))
  # => 'Hello @bob, say hi to @everyone in #general'
Slack::Messages::Formatting.unescape('Hello <@U123|bob> &gt; file.txt'))
  # => 'Hello @bob > file.txt'
Slack::Messages::Formatting.unescape('“hello”'))
  # => '"hello"'
Slack::Messages::Formatting.unescape('‘hello’'))
  # => "'hello'"

The slack command-line client returns JSON data from the Slack API.

$ slack --slack-api-token=[token] auth test
{"ok":true,"url":"...","team":"...","user":"...","team_id":"...","user_id":"..."}

export SLACK_API_TOKEN=...
$ slack chat postMessage --text="hello world" --channel="#general"
{"ok":true,"channel":"...","ts":"...","message":{"text":"hello world","username":"bot","type":"message","subtype":"bot_message","ts":"..."}}

$ slack channels id --channel=#general
{"ok":true,"channel":{"id":"C04KB5X4D"}}

$ slack channels info --channel=#general
{"ok":true,"channel":{"id":"C04KB5X4D","name":"general", ...}}

Combine with jq, a command-line JSON parser.

$ slack users list | jq '.members | map({(.id): .name})'
[
  {
    "U04KB5WQR": "dblock"
  },
  {
    "U07518DTL": "rubybot"
  }
]

See slack help for a complete command-line reference.

This gem is based on slack-ruby-gem, but it more clearly separates the Web and RTM APIs, is more thoroughly tested and is in active development.

See CONTRIBUTING.

Copyright (c) 2015-2019, Daniel Doubrovkine, Artsy and Contributors.

This project is licensed under the MIT License.