• Prerequisites
• Create the Python Files
• Install the LlamaIndex Packages
• Run the Python Script

LlamaIndex is a framework for building LLM-powered applications that use your own data.

Prerequisites

Before you can configure and use LlamaIndex with Connect AI, you must do the following:

• Connect a data source to your Connect AI account. See Sources for more information.

• Settings ページでPersonal Access Token（PAT）を生成します。PAT をコピーし、認証時にパスワードとして使用します。 OAuth JWT bearer token を生成します。PAT をコピーし、認証時にパスワードとして使用します。

• Obtain an OpenAI API key: https://platform.openai.com.

• Make sure you have Python >= 3.10 in order to install the LlamaIndex packages.

Create the Python Files

1. Create a folder for LlamaIndex MCP.

2. Create a Python file within the folder called llamaindex.py.

3. In llamaindex.py, set up your MCP server and MCP client to call the tools and prompts. For Authorization, you need to provide your Base64-encoded Connect AI username and PAT (obtained in the prerequisites) after Basic.For Authorization, you need to change Basic to Bearer and provide the JWT bearer token from the prerequisites.

from llama_index.llms.openai import OpenAI
from llama_index.tools.mcp import BasicMCPClient, aget_tools_from_mcp_url
from llama_index.core.agent import ReActAgent
import asyncio

client = BasicMCPClient(
    "http://mcp.cloud.cdata.com/mcp",
    headers={"Authorization": "Basic Base64-encoded (CONNECTAI_USERNAME:PAT)"}
)

async def main():
    # List available tools
    tools = await aget_tools_from_mcp_url("http://mcp.cloud.cdata.com/mcp", client=client)
   
    llm = OpenAI(model="gpt-4o", api_key="YOUR_OPENAI_KEY")   
    # Create ReActAgent
    agent = ReActAgent(tools=tools, llm=llm, verbose=True)
    # # Run a query
    response = await agent.run("List all the catalogs for me please")
    print(response)

asyncio.run(main())

Install the LlamaIndex Packages

Run pip install llama-index llama-index-llms-openai llama-index-tools-mcp in your project terminal.

Run the Python Script

1. When the installation finishes, run python llamaindex.py to execute the script.

2. The script discovers the Connect AI MCP tools needed for the LLM to query the connected data.

3. Supply a prompt for the agent. The agent provides a response.