Last released Jun 11, 2026
Code-mixed-aware prompt-injection guardrail for Indian voice agents — catches attacks in Hinglish/Indic speech without blocking genuine callers.
Supported by
