CrewAI Integration
CrewAI is a framework for orchestrating role-playing, autonomous AI agents. This guide shows how to integrate CrewAI with Cequence AI Gateway using the MCP (Model Context Protocol) adapter.
Requirements
- Python 3.8 or higher
- CrewAI
- crewai-tools
Install the required dependencies:
pip install crewai crewai-tools
Configuration
To integrate CrewAI with Cequence AI Gateway, you'll use the MCPServerAdapter from the crewai-tools library:
from crewai import Agent, Task, Crew, Process
from crewai_tools import MCPServerAdapter
# Configure the MCP server parameters for AI Gateway
server_params = {
"url": "<CEQUENCE_AI_GATEWAY_MCP_URL>", # Replace with your AI Gateway MCP URL
"transport": "streamable-http"
}
try:
with MCPServerAdapter(server_params) as tools:
print(f"Available tools from AI Gateway: {[tool.name for tool in tools]}")
# Create an agent with AI Gateway tools
gateway_agent = Agent(
role="AI Gateway Service Integrator",
goal="Utilize tools from Cequence AI Gateway via MCP.",
backstory="An AI agent specialized in leveraging enterprise AI gateway capabilities.",
tools=tools,
verbose=True,
)
# Define a task for the agent
gateway_task = Task(
description="Process requests using AI Gateway tools and capabilities.",
expected_output="Results from AI Gateway processing.",
agent=gateway_agent,
)
# Create a crew with the agent and task
gateway_crew = Crew(
agents=[gateway_agent],
tasks=[gateway_task],
verbose=True,
process=Process.sequential
)
# Execute the crew
result = gateway_crew.kickoff()
print("\nCrew Task Result:\n", result)
except Exception as e:
print(f"Error connecting to AI Gateway MCP server: {e}")
print("Ensure the AI Gateway MCP server is running and accessible at the specified URL.")
Key Components
MCPServerAdapter
The MCPServerAdapter
allows CrewAI agents to connect to MCP (Model Context Protocol) servers, including Cequence AI Gateway endpoints.
Configuration Options
- url: Your AI Gateway MCP endpoint URL
- transport: Use "streamable-http" for HTTP-based communication
Agent Creation
Create CrewAI agents that can utilize tools and capabilities provided by AI Gateway through the MCP interface.
Crew Orchestration
Use CrewAI's Crew
class to orchestrate multiple agents working together with AI Gateway tools.
Best Practices
- Authentication: Configure proper authentication for your AI Gateway connection
- Error Handling: Implement comprehensive error handling for network connections and tool execution
- Resource Management: Use context managers (
with
statements) to ensure proper resource cleanup - Monitoring: Leverage AI Gateway's monitoring capabilities to track agent performance
- Security: Follow security best practices when configuring access to AI Gateway
Advanced Usage
Multiple Agents with Shared Tools
from crewai import Agent, Task, Crew, Process
from crewai_tools import MCPServerAdapter
server_params = {
"url": "<CEQUENCE_AI_GATEWAY_MCP_URL>",
"transport": "streamable-http"
}
with MCPServerAdapter(server_params) as tools:
# Create multiple agents with access to AI Gateway tools
researcher = Agent(
role="Research Analyst",
goal="Gather and analyze data using AI Gateway capabilities.",
tools=tools,
verbose=True
)
writer = Agent(
role="Content Writer",
goal="Create content based on research findings.",
tools=tools,
verbose=True
)
# Define tasks for each agent
research_task = Task(
description="Research market trends using AI Gateway tools.",
agent=researcher
)
writing_task = Task(
description="Write a report based on research findings.",
agent=writer
)
# Create and execute the crew
crew = Crew(
agents=[researcher, writer],
tasks=[research_task, writing_task],
process=Process.sequential
)
result = crew.kickoff()
Repository Reference
For more information about CrewAI and MCP integration, see the CrewAI documentation.
Next Steps
- Configure monitoring and logging for your AI Gateway integration
- Explore advanced CrewAI features for complex multi-agent workflows
- Set up authentication and security for your AI Gateway connection
- Implement custom tools and capabilities through the MCP interface