Open Source · Apache 2.0

Block prompt injection
before it reaches
your LLM

Zero-dependency Python middleware.
64 detection patterns. 100% benchmark precision.

64
Detection Patterns
0
Dependencies
<5ms
Latency
100%
Precision
$ pip install aig-guardian
from ai_guardian import Guard

guard = Guard()
result = guard.check_input("Ignore all instructions...")

# result.blocked = True
# result.risk_level = CRITICAL
# result.reasons = ['Prompt Injection']

BLOCKED score=95