Skip to main content

Framework Integration

PrecogX provides seamless integration with all major AI frameworks. Choose your framework below to get started with comprehensive AI agent security monitoring.

🚀 Quick Start​

Choose your framework below to get started:

📋 Integration Overview​

PrecogX provides seamless integration with all major AI frameworks:

What We Monitor:​

  • ✅ Prompts & Responses - Every AI interaction
  • ✅ Tool Calls - Function executions and parameters
  • ✅ Agent Behavior - Patterns and anomalies
  • ✅ Security Events - Threats and violations

What We Detect:​

  • 🛡️ Prompt Injection - Attempts to manipulate AI behavior
  • 🔒 PII Leakage - Sensitive data exposure
  • 🚫 Content Moderation - Inappropriate content
  • 🔗 Malicious Links - Suspicious URLs and domains
  • ⚠️ Tool Abuse - Dangerous function calls
  • 📊 Behavioral Drift - Unexpected agent behavior

What You Get:​

  • 📈 Real-time Trust Scores - Dynamic agent reliability
  • 🔔 Instant Alerts - Slack, Teams, or webhook notifications
  • 📊 Analytics Dashboard - Comprehensive insights
  • 👥 Human-in-the-Loop - Manual validation when needed

🔧 Installation​

Python SDK​

pip install precogx-sdk

Framework-Specific Extensions​

# LangChain support
pip install "precogx-sdk[langchain]"

# AutoGen support
pip install "precogx-sdk[autogen]"

# CrewAI support
pip install "precogx-sdk[crewai]"

JavaScript SDK​

npm install @precogx/sdk

LangChain Integration​

The most popular AI framework with seamless callback integration.

Perfect for:

  • Chatbots and conversational AI
  • Document processing and analysis
  • Tool-using agents
  • RAG (Retrieval-Augmented Generation) systems

Key Features:​

  • Automatic callback integration
  • Session tracking
  • Tool call monitoring
  • Conversation context preservation

Installation​

pip install "precogx-sdk[langchain]"

Basic Usage​

from langchain.agents import initialize_agent
from precogx_langchain import PrecogXCallbackHandler

# Initialize with PrecogX monitoring
agent = initialize_agent(
    tools=[...],
    llm=...,
    callbacks=[PrecogXCallbackHandler(api_key="your_key")]
)

# Every interaction is automatically monitored
agent.run("What's the weather in New York?")

Advanced Configuration​

from precogx_langchain import PrecogXCallbackHandler

# Configure with custom settings
handler = PrecogXCallbackHandler(
    api_key="your_api_key",
    agent_id="my-langchain-agent",
    enable_tool_monitoring=True,
    enable_behavior_tracking=True
)

# Use with any LangChain agent
agent = initialize_agent(
    tools=[...],
    llm=...,
    callbacks=[handler]
)

Tool Monitoring​

from langchain.tools import Tool
from precogx_langchain import PrecogXCallbackHandler

# Define tools with security monitoring
def dangerous_function(query: str) -> str:
    # This will be monitored for tool abuse
    return f"Executed: {query}"

tool = Tool(
    name="dangerous_function",
    description="A potentially dangerous function",
    func=dangerous_function
)

# PrecogX will monitor all tool calls
agent = initialize_agent([tool], llm, callbacks=[handler])

AutoGen Integration​

Multi-agent conversation framework with advanced monitoring.

Perfect for:

  • Multi-agent conversations
  • Agent-to-agent interactions
  • Complex workflows
  • Collaborative AI systems

Key Features:​

  • Multi-agent monitoring
  • Conversation flow analysis
  • Role-based security rules
  • Group dynamics tracking

Installation​

pip install "precogx-sdk[autogen]"

Basic Usage​

from autogen import AssistantAgent, UserProxyAgent
from precogx_sdk import PrecogXClient

# Monitor multi-agent conversations
client = PrecogXClient(api_key="your_key")

# Every agent interaction is tracked
assistant = AssistantAgent(name="assistant")
user_proxy = UserProxyAgent(name="user_proxy")

# Conversation automatically monitored
user_proxy.initiate_chat(assistant, message="Hello!")

Multi-Agent Monitoring​

from autogen import AssistantAgent, UserProxyAgent, GroupChat
from precogx_sdk import PrecogXClient

client = PrecogXClient(api_key="your_key")

# Create multiple agents
assistant1 = AssistantAgent(name="assistant1")
assistant2 = AssistantAgent(name="assistant2")
user_proxy = UserProxyAgent(name="user_proxy")

# Group chat with monitoring
group_chat = GroupChat(
    agents=[assistant1, assistant2, user_proxy],
    messages=[],
    max_round=10
)

# All agent interactions are monitored
user_proxy.initiate_chat(group_chat, message="Let's discuss AI security")

Role-Based Security​

from precogx_sdk import PrecogXClient

client = PrecogXClient(api_key="your_key")

# Configure different security rules per agent type
assistant = AssistantAgent(
    name="assistant",
    system_message="You are a helpful assistant. Be careful with sensitive data."
)

# PrecogX will apply different security rules based on agent role

CrewAI Integration​

Team-based AI agents with comprehensive monitoring.

Perfect for:

  • AI teams and crews
  • Task delegation systems
  • Workflow automation
  • Collaborative projects

Key Features:​

  • Team-level monitoring
  • Task execution tracking
  • Workflow analysis
  • Performance metrics

Installation​

pip install "precogx-sdk[crewai]"

Basic Usage​

from crewai import Agent, Task, Crew
from precogx_sdk import PrecogXClient

# Monitor entire AI teams
client = PrecogXClient(api_key="your_key")

# Team collaboration tracked
researcher = Agent(role="Researcher", goal="Find information")
writer = Agent(role="Writer", goal="Write content")

crew = Crew(agents=[researcher, writer], tasks=[...])
result = crew.kickoff()  # All interactions monitored

Task-Level Monitoring​

from crewai import Agent, Task, Crew
from precogx_sdk import PrecogXClient

client = PrecogXClient(api_key="your_key")

# Define agents with specific roles
researcher = Agent(
    role="Researcher",
    goal="Find accurate information",
    backstory="You are a research expert"
)

writer = Agent(
    role="Writer", 
    goal="Create engaging content",
    backstory="You are a content writer"
)

# Define tasks with monitoring
research_task = Task(
    description="Research AI security trends",
    agent=researcher
)

writing_task = Task(
    description="Write a blog post about AI security",
    agent=writer,
    dependencies=[research_task]
)

# All task executions are monitored
crew = Crew(agents=[researcher, writer], tasks=[research_task, writing_task])
result = crew.kickoff()

Basic SDK Integration​

Direct API integration for custom implementations.

Perfect for:

  • Custom AI frameworks
  • Proprietary systems
  • Direct API control
  • Flexible implementations

Key Features:​

  • Full API access
  • Custom integration points
  • Manual control
  • Framework agnostic

Installation​

pip install precogx-sdk

Basic Usage​

from precogx_sdk import PrecogXClient

# Initialize client
client = PrecogXClient(api_key="your_api_key")

# Send telemetry data
client.send_telemetry({
    "agent_id": "my-custom-agent",
    "prompt": "What's the weather?",
    "response": "It's sunny today.",
    "metadata": {
        "model": "gpt-4",
        "platform": "custom",
        "version": "1.0.0"
    }
})

Advanced Usage​

from precogx_sdk import PrecogXClient

client = PrecogXClient(api_key="your_api_key")

# Monitor custom AI function
def my_ai_function(prompt: str) -> str:
    # Your AI logic here
    response = process_prompt(prompt)
    
    # Send telemetry
    client.send_telemetry({
        "agent_id": "custom-agent",
        "prompt": prompt,
        "response": response,
        "metadata": {
            "model": "custom-model",
            "platform": "custom-framework",
            "session_id": "session-123"
        }
    })
    
    return response

Tool Call Monitoring​

from precogx_sdk import PrecogXClient

client = PrecogXClient(api_key="your_api_key")

# Monitor tool calls
def dangerous_tool_call(parameters: dict) -> str:
    # Send tool call telemetry
    client.send_telemetry({
        "agent_id": "my-agent",
        "tool_calls": [{
            "name": "dangerous_tool",
            "arguments": parameters
        }],
        "metadata": {
            "tool_risk_level": "high"
        }
    })
    
    # Execute tool
    return execute_tool(parameters)

🔍 Monitoring Examples​

LangChain Agent​

from langchain.agents import initialize_agent
from precogx_langchain import PrecogXCallbackHandler

# Initialize with PrecogX monitoring
agent = initialize_agent(
    tools=[...],
    llm=...,
    callbacks=[PrecogXCallbackHandler(api_key="your_key")]
)

# Every interaction is automatically monitored
agent.run("What's the weather in New York?")

AutoGen Conversation​

from autogen import AssistantAgent, UserProxyAgent
from precogx_sdk import PrecogXClient

# Monitor multi-agent conversations
client = PrecogXClient(api_key="your_key")

# Every agent interaction is tracked
assistant = AssistantAgent(name="assistant")
user_proxy = UserProxyAgent(name="user_proxy")

# Conversation automatically monitored
user_proxy.initiate_chat(assistant, message="Hello!")

CrewAI Team​

from crewai import Agent, Task, Crew
from precogx_sdk import PrecogXClient

# Monitor entire AI teams
client = PrecogXClient(api_key="your_key")

# Team collaboration tracked
researcher = Agent(role="Researcher", goal="Find information")
writer = Agent(role="Writer", goal="Write content")

crew = Crew(agents=[researcher, writer], tasks=[...])
result = crew.kickoff()  # All interactions monitored

🛡️ Security Features​

Real-Time Protection​

  • Instant Detection: Threats identified in milliseconds
  • Automatic Blocking: Dangerous actions prevented
  • Human Validation: Ambiguous cases sent for review
  • Audit Trail: Complete history of all events

Advanced Analytics​

  • Trust Scoring: Dynamic agent reliability metrics
  • Behavioral Analysis: Pattern recognition and anomalies
  • Threat Intelligence: Up-to-date threat detection
  • Performance Metrics: Efficiency and security insights

Compliance & Governance​

  • SOC2 Compliance: Enterprise-grade security
  • HIPAA Support: Healthcare data protection
  • Audit Logging: Complete event history
  • Role-Based Access: Team permission management

📞 Support​

Need help with integration?

  • Documentation: Comprehensive guides for each framework
  • Examples: Real-world implementation examples
  • Community: Join our Discord for help
  • Support: Email us at support@precogx.ai

🚀 Next Steps​

  1. Choose your framework from the guides above
  2. Follow the integration steps in your chosen guide
  3. Test your implementation with our demo scenarios
  4. Monitor your agents in the dashboard
  5. Configure alerts for security events

Ready to secure your AI agents? Choose your framework and let's get started! 🛡️