Framework Support
Works With Your Stack
LangChain
Native callbacks for seamless integration
LlamaIndex
Built-in support for RAG pipelines
AutoGen
Multi-agent orchestration support
CrewAI
Agent crew monitoring & control
Google ADK
First-class ADK integration
OpenAI SDK
Direct API interception
Simple Integration
Add Security in Minutes
Just a few lines of code to protect your entire agent pipeline. Our SDK wraps your existing code without changing your architecture.
- OpenTelemetry-native tracing
- Real-time threat scanning
- Automatic cost tracking
- Compliance audit logging
from saf3ai_sdk import init, create_security_callback
# Initialize SDK
init(
service_name="my-agent",
framework="langchain",
agent_id="agent-001",
)
# Create security callback
security_cb = create_security_callback(
api_endpoint="https://api.saf3ai.com",
)
# Add to your LangChain agent
chat = ChatOpenAI(callbacks=[security_cb])Capabilities
Everything You Need for Production
Security Scanning
Real-time detection of prompt injections, jailbreaks, and malicious inputs.
Observability
Full tracing of agent workflows with latency, errors, and performance metrics.
Cost Tracking
Monitor token usage and costs across all providers with granular attribution.
Compliance
Audit logs and reporting ready for SOC 2, GDPR, HIPAA requirements.
Evaluations
Test agent quality, safety, and performance before deploying to production.
Alerting
Get notified immediately when security events or anomalies are detected.
Ready to secure your agents?
Get started in minutes with our SDK and comprehensive documentation.