AI content moderation uses AI to classify incoming text so unwanted messages can be handled before they become a distraction. In Vibebo, AI Sentinel works with your own safety settings to help detect spam or promotional messages, harmful content and sexually suggestive content sent through your anonymous message link or Channels.
What AI content moderation looks for
AI Sentinel looks at the meaning and context of a message rather than relying only on a fixed list of banned words. Depending on the filters you enable, Vibebo can identify messages that appear to be spam or promotional, harmful or abusive, or sexually suggestive.
How AI Sentinel works with your inbox
The public AI Sentinel experience lets you test how Vibebo classifies text. Your own message filtering is managed from Settings > Safety > Message Filtering. That is where you decide which AI-assisted filters you want to use on messages sent to your profile.
Turn message filtering on or off
- Open Vibebo and go to Settings.
- Open Safety.
- Choose Message Filtering.
- Turn the available filtering categories on or off to match the level of filtering you want.
- Return to your inbox and adjust the settings later if your needs change.
Use Hidden Words for your own boundaries
AI filtering and Hidden Words solve different problems. AI Sentinel classifies broader message types such as spam, harmful content and sexually suggestive content. Hidden Words lets you add specific words, phrases or emoji that you personally do not want in your inbox. Using both gives you a general moderation layer plus your own custom boundaries.
Test AI moderation with AI Sentinel
You can open the AI Sentinel demo to try sample text and see how the moderation experience classifies it. The demo is useful when you want to understand the feature before deciding which filtering options to enable on your own account.
Use more than one safety control
Message filtering is one part of Vibebo’s safety controls. If a message still reaches you and crosses a boundary, you can use the message actions available in Inbox+ to block the sender or report the content. Hidden Words can also be updated whenever a new phrase or topic becomes something you no longer want to receive.
AI helps with moderation, but language can be ambiguous and context matters. Review the controls that work best for your audience and combine automated filtering with the boundaries you set yourself.
What happens after you change a filter
Once you change Message Filtering, use your inbox normally and pay attention to whether the settings match the experience you want. Different audiences can produce very different kinds of messages, so a setup that works for a private Channel may not be the same setup you want for a personal link shared widely.
If legitimate messages appear to be caught too often, review the categories you have enabled and check your Hidden Words list for terms that may be too broad. If unwanted messages keep appearing, strengthen the relevant filtering category or add a recurring term to Hidden Words.
Choose between AI filtering and Hidden Words
Use AI-assisted filtering when the problem is a broader type of content. Spam can be written in many different ways; harmful or sexually suggestive messages can also avoid any single obvious keyword. Sentinel is intended to classify the overall message.
Use Hidden Words when you already know the exact wording you want to exclude. This is useful for a specific phrase, nickname, slur, repeated topic or emoji that matters to your own space.
Common questions about message filtering
Can I turn filtering off later?
Yes. Return to Settings > Safety > Message Filtering and change the categories you have enabled.
Does using AI filtering reveal who sent a message?
No. Message filtering is about the content being submitted. The sender remains anonymous to you.
Do I need Hidden Words if AI filtering is on?
Not necessarily, but the two controls do different jobs. Hidden Words is useful for personal boundaries that may not fit a broad moderation category.
What should I do with a serious message that reaches Inbox+?
Use the actions available with the message. You can report content that should be reviewed and block a sender when you want to stop further interaction from that sender.
A good filtering setup is one you can understand and adjust. Start with the categories that match the problem you are actually seeing, add specific Hidden Words where necessary, and review the settings when your audience or use case changes.