AI Malware Filter
Machine learning based malicious domain classifier.
Overview
This is an experimental malware filter that uses machine learning to classify domains as malicious, without them appearing in any block list.
Usage Warning
This feature is experimental and may cause false positive blocks for less popular websites, on uncommon TLDs. Use at own risk.
Modes
You can choose one of 3 different modes, which will dictate how aggressively the Filter will behave.
- Relaxed - Only blocks domains that have a very high confidence that they are malicious. This setting is recommended.
- Balanced - Will block domains that have a medium confidence that they are malicious.
- Strict - Will block domains that have a medium/low confidence that they are malicious. This is not recommended as there will be a lot of false positives.
False Positives
As this feature is still experimental and undergoing rapid development, we do expect there to be some false positives. If a domain is blocked by the AI Malware filter and you believe it shouldn't be, we suggest two courses of action:
- Popular Domains - Report the false positive to us at [email protected]
- Lesser-known Domains - Create a Custom Rule to Bypass the domain, which you can do from your Profile or by using the handy pencil icon button in the Activity Log
Updated 11 months ago