Start securing AI in 60 seconds
Free forever. Full detection engine. 515+ signatures. 100% local. No API key required.
Three steps to runtime AI security
Install
pip install raxeIntegrate
from raxe import scanner
@scanner.protect
def ask_ai(prompt: str) -> str:
return openai.chat(prompt)Observe
[RAXE] Threat detected: prompt_injection
severity: critical
technique: AML.T0051 (Direct Prompt Injection)
action: blocked
latency: 0.12msWorks with your AI framework
LangChain
Chain & agent protection
CrewAI
Multi-agent security
AutoGen
Conversation guardrails
LiteLLM
Provider-agnostic scanning
LlamaIndex
RAG pipeline security
DSPy
Program-level detection
Anthropic
Claude wrapper
Portkey
Gateway integration
Test detection quality in your browser
100% local. No installation required. Try the full detection engine with real attack patterns.
Try the Live ScannerWhat you get
5 Detection Layers
Pattern Rules, ML Model, Statistics, SLM Judge, Memory
515+ Signatures
MITRE ATLAS mapped, open YAML detection rules
On-Device ML
ONNX inference, no GPU required, 100% local
<10ms P95 Latency
Sub-millisecond pattern matching, lightweight ML
14 Threat Families
35 attack techniques, 10 harm types
Log-Only Mode
Safe initial deployment, observe before enforce
pip install
Pod-level
Node-level
Network-level
When your team is ready for infrastructure-wide AI security, the platform is ready for you.