Content moderation overview
The IM service supports content moderation for IM messages via a built-in sensitive word mechanism. Note that this mechanism is a basic protection measure, applies only to text content (up to 50 sensitive words by default), and is not a replacement for a professional content moderation service.
If a message is blocked due to a violation and cannot be delivered to the recipient, the message sender is not notified by default. See Sensitive Content Interception Callback.
Sensitive word mechanism
Note
The client does not provide a management interface for this feature. It only provides a callback interface to notify the sender when a message is blocked. See Sensitive Content Interception Callback.
The sensitive word mechanism is a basic protection measure that supports detecting and filtering sensitive words in text message content only. For messages that match sensitive words, you can choose to block the message (not delivered to the recipient) or replace the sensitive words according to specified rules before delivering it.
Supported languages for sensitive word filtering: Chinese, English, Japanese, German, Russian, Korean, Arabic.
Manage sensitive words for the development or production environment under an App Key via the Server API:
| Feature | Server API |
|---|---|
| Add sensitive words (with replacement content) | Add Sensitive Word |
| Remove a sensitive word | Remove Sensitive Word |
| Batch remove sensitive words | Batch Remove Sensitive Words |
| Get sensitive word list (with replacement content) | Get Sensitive Word List |
Default behavior
- Up to 50 sensitive words by default.
- Applies only to messages sent from the client SDK by default.
- By default, only detects sensitive words in the built-in text message type (message identifier
RC:TxtMsg). Supported Channel types: direct, group, open channel, and community channel. Modified text message content in super groups is also subject to sensitive word detection, blocking, and filtering by default.
Adjust configuration
- To filter sensitive words in messages sent via the server API, enable the Moderate messages sent by server API toggle under Chat > Chat settings > Security & Moderation in the console.