twitch-bot/wiki/Home.md

Configuration

---

# Channels to join (only those can be acted on)
channels:
  - mychannel

# Allow moderators to hand out permits (if set to false only broadcaster can do this)
permit_allow_moderator: true
# How long to permit on !permit command
permit_timeout: 60s

rules: # See below for examples

  - actions: # Array of actions to take when this rule matches

    # Issue a ban on the user who wrote the chat-line
    - ban: "reason of ban"

    # Command to execute for the chat message, must return an JSON encoded array of actions
    - command: [/bin/bash, -c, "echo '[{\"respond\": \"Text\"}]'"]

    # Modify an internal counter value (does NOT send a chat line)
    - counter: "counterid" # String to identify the counter, applies templating
      counter_step: 1      # Integer, can be negative or positive, default: +1

    # Issue a delete on the message caught
    - delete_message: true # Bool, set to true to delete

    # Send responding message to the channel the original message was received in
    - respond: 'Hello chatter' # String, applies templating

    # Issue a timeout on the user who wrote the chat-line
    - timeout: 1s # Duration value: 1s / 1m / 1h

    # Add a cooldown to the command (not to trigger counters twice, ...)
    cooldown: 1s # Duration value: 1s / 1m / 1h

    # Disable actions on this rule if the user has an active permit
    disable_on_permit: false

    # Disable actions on this rule if the user has one of these badges
    disable_on: [broadcaster, moderator]

    # Enable actions on this rule only if the user has one of these badges
    enable_on: [broadcaster, moderator]

    # Require the chat message to be sent in this channel
    match_channels: ['#mychannel']

    # Require the chat message to be sent by one of these users
    match_users: ['mychannel'] # List of users, all names MUST be all lower-case

    # Execute actions when this event occurs
    match_event: 'permit'

    # Execute action when the chat message matches this regular expression
    match_message: '' # String, regular expression

    # Disable the actions on this rule if one of these regular expression matches the chat message
    disable_on_match_messages: []

...

Templating

There are certain variables available in the strings with templating enabled:

• channel - Channel the message was sent to, only available for regular messages not events
• msg - The message object, used in functions, should not be sent to chat
• permitTimeout - Value of permit_timeout in seconds
• username - The username of the message author

Additionally there are some functions available in the templates:

• arg <idx> - Takes the message sent to the channel, splits by space and returns the Nth element
• channelCounter <counter name> - Wraps the counter name into a channel specific counter name including the channel name
• counterValue <counter name> - Returns the current value of the counter which identifier was supplied
• fixUsername <username> - Ensures the username no longer contains the @ prefix
• group <idx> - Gets matching group specified by index from match_message regular expression
• recentGame <username> [fallback] - Returns the last played game name of the specified user (see shoutout example) or the fallback if the game could not be fetched. If no fallback was supplied the message will fail and not be sent.
• tag <tagname> - Takes the message sent to the channel, returns the value of the tag specified

Command executions

Your command will get a JSON object passed through stdin you can parse to gain details about the message. It is expected to yield an array of actions on stdout and exit with status 0. If it does not the action will be marked failed. In case you need to output debug output you can use stderr which is directly piped to the bots stderr.

This is an example input you might get on stdin:

{
  "badges": {
    "glhf-pledge": 1,
    "moderator": 1
  },
  "channel": "#tezrian",
  "message": "!test",
  "tags": {
    "badge-info": "",
    "badges": "moderator/1,glhf-pledge/1",
    "client-nonce": "6801c82a341f728dbbaad87ef30eae49",
    "color": "#A72920",
    "display-name": "Luziferus",
    "emotes": "",
    "flags": "",
    "id": "dca06466-3741-4b22-8339-4cb5b07a02cc",
    "mod": "1",
    "room-id": "485884564",
    "subscriber": "0",
    "tmi-sent-ts": "1610313040489",
    "turbo": "0",
    "user-id": "69699328",
    "user-type": "mod"
  },
  "username": "luziferus"
}

The example was dumped using this action:

  - actions:
    - command: [/usr/bin/bash, -c, "jq . >&2"]
    match_channels: ['#tezrian']
    match_message: '^!test'

Rule examples

Game death counter with dynamic name

  - actions:
    - counter: '{{ channelCounter (recentGame .channel) }}'
    - respond: >-
        I already died {{ counterValue (channelCounter (recentGame .channel)) }}
        times in {{ recentGame .channel }}'        
    cooldown: 60s
    enable_on: [broadcaster, moderator]
    match_channels: ['#mychannel']
    match_message: '^!death'

Link-protection while allowing Twitch clips

  - actions:
    - timeout: 1s
    - respond: '@{{ .username }}, please ask for permission before posting links.'
    disable_on: [broadcaster, moderator, subscriber, vip]
    disable_on_match_messages:
      - '^(?:https?://)?clips\.twitch\.tv/[a-zA-Z0-9-]+$'
    disable_on_permit: true
    match_channels: ['#mychannel']
    match_message: '(?:[a-z0-9](?:[a-z0-9-]{0,61}[a-z0-9])?\.)+[a-z0-9][a-z0-9-]{0,61}[a-z0-9]'

Send a notification on successful permit

  - actions:
    - respond: >-
        @{{ fixUsername (arg 1) }}, you will not get timed out
        for the next {{ .permitTimeout }} seconds.        
    match_channels: ['#mychannel']
    match_event: 'permit'

Shoutout command with game query

  - actions:
    - respond: >-
        Check out @{{ fixUsername (group 1) }} and leave a follow,
        they were last playing {{ recentGame (fixUsername (group 1)) "something mysterious" }}
        at https://twitch.tv/{{ fixUsername (group 1) }}        
    enable_on: [broadcaster, moderator]
    match_channels: ['#mychannel']
    match_message: '^!so ([@\w]+)'