Observed arrival · 2026-11-08
Maskena checks what your AI can’t see
A Portuguese-language tool and API that scan files for hidden instructions aimed at AI systems and redact personal data before documents reach a model or RAG pipeline.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- AI users handling sensitive documents
- Worth noticing
- The site describes three detection layers, including rules in ten languages and Llama Prompt Guard 2 in its highest analysis tier.
Field notes
The homepage distinguishes hidden text—such as white or microscopic lettering and invisible characters—from personal-data redaction. It also describes checks for external links, password requests, and agent commands such as terminal actions or downloads. The site says files retain their original format after cleanup, while scanned images and PDFs receive black bars over detected personal data.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
$PaidCommerce or pricing visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue