Context Detection (Beta)

Context Detection is a next-generation AI protection layer that analyzes the meaning and intent behind messages — not just individual keywords. Unlike traditional filters that rely on exact word matches, Context Detection evaluates the full context of phrases to identify violations that might otherwise go unnoticed.

How It Works

Context Detection functions as a global filter, operating independently from other Message Guard layers, including:

• Red/Yellow word lists,

• Custom lists,

• Bypass Protection.

The system uses a proprietary detection algorithm that:

1. Identifies messages containing trigger words from an internal monitoring list

2. Analyzes the full context and intent of the message

3. Flags violations based on meaning, even when no explicitly prohibited words are used

Example Detection:

• Message: "Maybe we could chat somewhere else?"

• Result: Flagged as a Direct Contact violation even though it contains no explicitly banned terms.

This layer is particularly effective at detecting indirect or conversational phrasing used to bypass keyword-based filters.

Protected Topics

Context Detection monitors the following four violation categories:

• Direct Contact - Attempts to move the conversation to external messaging platforms

• Real-world Meetup Escort - Proposals for physical meetings or encounters

• Third-party Payment - Attempts to redirect payments outside the platform

• Underage - Any content or language related to minors.

Important Notes:

• These four categories can be disabled during the beta phase.

• Context Detection works in addition to your existing word filters, providing an extra layer of protection

• All violations are logged in the standard Message Guard Logs, with the topic clearly identified.

Relationship to Other Protection Layers

Context Detection is designed to work alongside your existing Message Guard configuration:

• Independent Operation - It functions separately from Red/Yellow/Custom word lists.

• Dual Protection - A message can be blocked by keyword filters, Context Detection or both.

• Unified Logging - All violations—regardless of detection method—appear together in the Message Guard Logs, clearly labeled with the appropriate detection source.

Summary

Context Detection adds an intelligent, meaning-based protection layer that adapts to real conversation patterns. You keep full control of your traditional word lists, while Context Detection enhances security by identifying violations based on intent, not just keywords.