Event detection

Event detection is the first step in the rule engine where events are automatically detected using the profanity filter or Hive AI. When detected, it alerts the engine to check if the event falls under the conditions set in the rule engine. Currently, only one detection method—either the profanity filter or Hive AI—can be selected at a time as they cannot be operated simultaneously.

Note: Hive AI is an add-on feature. Contact us if you wish to use this feature.

Before you begin, the Moderation dashboard must be enabled. You can do so under your Application > Settings > Moderation > Features.

To view the image in full screen, click on the

Before setting up the rule engine, you need to set up the profanity filter. The profanity filter is a feature that automatically detects and filters out abusive language in messages by defining offensive words and regular expression patterns.

Note: The global filter applies ot all channels, regardless of the channel's custom type. You can also create filters for specific custom channel types and decide whether to include or exclude the global filter in these channels.

1. On the Sendbird Dashboard, go to your Application > Settings > Moderations > Profanity filter.

To view the image in full screen, click on the

2. Turn the Filter on.
3. Select one of the following options:
   • No action: Message is sent to channel without any changes and is automatically forwarded to Rule engine for further moderation.
   • Replace with asterisks(*): The offensive word in a message is replaced by asterisks and sent to the channel.
   • Block from sending: The message containing an offensive word is not sent to the channel.
4. Then add words to the Offensive words list to be filtered and detected as profanity,
5. You can also specify regular expression patterns to be detected as profanity.
6. Decide if you choose to include or exclude the global filter in this specific custom channel types.

Hive AI enhances content moderation by automating the detection of complex behaviors and content that may not be explicitly profane but still violates guidelines. This integration allows for more nuanced and effective moderation, maintaining a clean and safe environment for user interactions. To utilize Hive AI, a no-code integration that does not require additional development, follow the steps below.

Note: Hive AI is an add-on feature. Contact us if you wish to use this feature.

1. On the Sendbird Dashboard, go to your Application > Settings > Moderations > Integrations > Hive.

To view the image in full screen, click on the

2. Choose an automated response in the Action section. You can select from the following message delivery modes:

3. Set a severity level limit for each category of content. When message content meets or exceed the specified level, it will automatically initiate the action response in the Action section.

You can also set up keyword detection. The keyword detection is a feature that automatically detects and filters out abusive language in messages by defining offensive words and regular expression patterns. You can do so under Application > Settings > Moderation > Advanced moderation > Keyword detection on the Sendbird Dashboard. To learn more, refer to the Keyword detection page.

Note: This is an Advanced moderation feature. If you wish to use this feature, contact sales.

To view the image in full screen, click on the

Explore more on our Advanced moderation product page.