Mattral/GuardRail-Studio
Ultra-Low-Latency, High-Throughput LLM Firewall & Observability Platform Designed to defend planet-scale LLM systems against prompt injection, PII leakage, data poisoning, and model drift — with a sub-10 millisecond p99 latency design target.
GitHub repository with 9 stars and 7 forks.
Language: Python