Content Guard
Checks LLM outputs and drafts for hidden steganographic leakage before you share them. Strips invisible Unicode, normalizes text, flags anomalies, and hands back a clean copy for your approval. It never sends anything itself.
What it does
- Checks LLM outputs
- Drafts for hidden steganographic leakage before you share them
- Flags anomalies
In the creator’s own words.