How to set up sensitive words? Configure built-in sensitive word service

[RC]

The Instant Messaging (IM) service includes a built-in sensitive word feature. This guide shows how to set it up in the Developer Console. Supported languages for filtering include: Chinese, English, Japanese, German, Russian, Korean, and Arabic.

The sensitive word feature is a basic protection tool for text only (default limit: 50 words). It doesn’t replace professional content moderation.

Steps

Go to the RC Console - Sensitive Words Settings page to:

• Add or remove sensitive words

• Choose actions: block messages or replace sensitive characters

• Export the sensitive word list

• Add custom message types for filtering (default: only works with RC’s built-in text messages, ObjectName: RC:TxtMsg)

Click Add on the Sensitive Words Settings page and select the tab to add words.

Notify senders when a message is blocked

By default, senders aren’t notified if a message is blocked due to sensitive words. If you want senders to know, submit a ticket to enable Blocked Message Status Callback to Sender and set up a listener on the client side (requires Android/iOS SDK version ≧ 5.1.4, Web ≧ 5.0.2).

See the “Message Block Notification Callback” docs for each client:

• Android: https://docs.rongcloud.io/android-imlib/moderation/sensitive

• iOS: https://docs.rongcloud.io/ios-imlib/moderation/sensitive

• Web: https://docs.rongcloud.io/web-imlib/moderation/sensitive

Enable sensitive word filtering for server API messages

By default, messages sent via the server API aren’t filtered. To enable it, go to the Basic Features section in the Developer Console and turn on Filter Server API messages.

Set up sensitive word filtering for custom messages

Click Set custom messages on the Sensitive Words Settings page. Provide the ObjectName of the custom message type and the key in the content JSON structure. Once set up, the value for this key will be filtered.

Notes on message identifiers:

1. All message identifiers (ObjectName) you add will filter the content of their corresponding key. For example, the built-in “Text Message” type (ObjectName: RC:TxtMsg) filters the value of the “content” key in the message body.

2. Custom messages must follow SDK standards and be valid JSON.

3. Avoid starting with “RC:” to prevent conflicts with RC’s built-in message ObjectNames.

4. Test in the development environment before deploying to production to ensure filtering works. If it fails, submit a ticket.

Adjust the sensitive word limit

Increasing the sensitive word limit is a paid service. The default is 50 words. To adjust, go to the Extended Services tab in the Chat Pricing Plans page in the Developer Console. Only Chat Premium Plan or Chat Ultimate Plan users can do this. See billing details for costs.

• Chat Premium Plan, Chat Ultimate Plan Billing (global)

References

How to get original message content when using built-in sensitive words, IM moderation, and message callback services?

Description Instant messaging supports IM moderation, offering: Built-in sensitive words: Automatically blocks or replaces sensitive words before delivery. Professional IM moderation: RC’s content moderation service moderates text, images, voice clips, and short videos to identify sensitive content. Message callback: Sends messages to your app first, letting you decide whether to deliver them. Use this to implement your own moderation or integrate third-party services. When moderating content…

Built-in sensitive word mechanism – blocked message vs replace sensitive word: what’s the difference in filtering effect?

Built-in sensitive word filtering methods Two methods are available: Blocked message: Messages containing sensitive words will be blocked by the RC server, and the target user won’t receive them. Replace sensitive word: Messages containing sensitive words will have those words replaced with specified content before being sent to the target user. If a message contains both types of sensitive words, it will be blocked. Retrieve the original content of messages containing sensitive word…

Understand RC's sensitive word filtering rules and matching effects

When using the built-in sensitive word mechanism in Instant Messaging (IM) services, you may need to add sensitive words in the developer console at How to set up sensitive words? Configure built-in sensitive word service. This article outlines the setup rules and matching effects for sensitive word strings. Simplified and Traditional Chinese smart filtering Setting Simplified Chinese sensitive words automatically filters the corresponding Traditional Chinese sensitive words. Smartly punctuati…

Need help?

Submit a ticket if you have questions.