Observed arrival · 2026-10-13
Saaf by Safearly: an India-focused PII firewall for LLMs
Saaf masks Indian personal identifiers before they reach an LLM, then restores the real values in the reply.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Teams sending Indian user data to LLMs
- Worth noticing
- Its published benchmark uses 420 synthetic documents and calls the results a smoke test, while reporting 0.83 false alarms per 1,000 tokens.
Field notes
The described pipeline normalizes hidden characters, lookalikes and Indic digits before checking candidate matches; policies can mask, block, warn or allow them. Saaf says real values remain in a short-lived vault on the customer’s infrastructure, while its benchmark currently covers 420 synthetic documents and is explicitly labeled a smoke test.
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