SDK installation, Monitor vs Proxy modes, multi-provider support, CI/CD integration, MCP Server, API versioning, and best practices.
AgentAIShield provides official SDKs for Node.js and Python with identical feature sets.
Node.js: npm install @agentaishield/sdk
Python: pip install agentaishield
from agentaishield import AgentAIShield
# Initialize the client
aais = AgentAIShield(api_key='aais_...')
# Monitor mode (fire-and-forget)
aais.monitor(
agent_id='chatbot-v1',
messages=[{'role': 'user', 'content': 'Hello!'}],
response={'content': 'Hi there!'}
)
# Proxy mode (inline scanning)
response = aais.proxy.chat.completions.create(
agent_id='chatbot-v1',
model='gpt-4',
messages=[{'role': 'user', 'content': 'Hello!'}]
)
const { AgentAIShield } = require('@agentaishield/sdk');
// Initialize the client
const aais = new AgentAIShield({ apiKey: 'aais_...' });
// Monitor mode (fire-and-forget)
await aais.monitor({
agentId: 'chatbot-v1',
messages: [{ role: 'user', content: 'Hello!' }],
response: { content: 'Hi there!' }
});
// Proxy mode (inline scanning)
const response = await aais.proxy.chat.completions.create({
agentId: 'chatbot-v1',
model: 'gpt-4',
messages: [{ role: 'user', content: 'Hello!' }]
});
AgentAIShield supports two authentication methods:
All trust verification and TrustShield endpoints accept BOTH authentication methods, making it seamless to use features from the dashboard without creating an API key first.
OpenAI, Anthropic, Google (Gemini), Mistral, Cohere, Azure OpenAI. Unified interface across all providers.
GitHub Actions, GitLab CI, CircleCI. Run security scans in your pipeline before deployment.