Skip to main content
This feature is essential for proactive impersonation detection and should be part of your moderation workflow.

​
How It Works

  • Hashbot scans your server for avatars similar to any image in your active filter list
  • It presents a list of suggested matches for manual review
  • You can choose to ban some, all, or none of the flagged users
These are near matches — they may not cross the automatic similarity threshold but still closely resemble known scam images.

​
Hashbot Learns from Your Input

Every time you confirm a match during PFP Review, Hashbot improves its detection.
  • Confirmed bans help Hashbot recognize future visual variations
  • Over time, this reduces reliance on exact image matches

​
Example

If a scammer flips your team logo or shifts it slightly, it may not trigger the automated threshold. But if you flag it manually, Hashbot will learn to detect similar distortions in the future. Assign PFP Review as a regular task to trusted moderators or security leads. When to run:
  • Once a week as part of routine review
  • After applying the /template pfp list
  • After uploading new team or brand-related images
  • In combination with /pfp-filters cleanse to catch users already in the server