Use Agno to Talk to Your Databricks Data via CData Connect AI

Agno is a developer-first Python framework for building AI agents that reason, plan, and take actions using tools. Agno emphasizes a clean, code-driven architecture where the agent runtime remains fully under developer control.

CData Connect AI provides a secure cloud-to-cloud interface for integrating hundreds of enterprise data sources with AI systems. Using Connect AI, live Databricks data data can be exposed through a remote MCP endpoint without replication.

In this guide, we build a production-ready Agno agent using the Agno Python SDK. The agent connects to CData Connect AI via MCP using streamable HTTP, dynamically discovers available tools, and invokes them to query live Databricks data.

Prerequisites

1. Python 3.9+.
2. A CData Connect AI account – Sign up or log in here.
3. An active Databricks account with valid credentials.
4. An LLM API key (for example, OpenAI).

Overview

Here is a high-level overview of the process:

1. Connect: Configure a Databricks connection in CData Connect AI.
2. Discover: Use MCP to dynamically retrieve tools exposed by CData Connect AI.
3. Query: Wrap MCP tools as Agno functions and query live Databricks data.

Getting Started

---

Step 1: Configure Databricks in CData Connect AI

To enable Agno to query live Databricks data, first create a Databricks connection in CData Connect AI. This connection is exposed through the CData Remote MCP Server.

1. Log into Connect AI, click Sources, and then click Add Connection. 👁 Adding a connection
   
2. Select "Databricks" from the Add Connection panel. 👁 Selecting a data source
   
3. Enter the required authentication properties.

   To connect to a Databricks cluster, set the properties as described below.

   Note: The needed values can be found in your Databricks instance by navigating to Clusters, and selecting the desired cluster, and selecting the JDBC/ODBC tab under Advanced Options.

   • Server: Set to the Server Hostname of your Databricks cluster.
   • HTTPPath: Set to the HTTP Path of your Databricks cluster.
   • Token: Set to your personal access token (this value can be obtained by navigating to the User Settings page of your Databricks instance and selecting the Access Tokens tab).
   👁 Configuring connection properties
   Click Create & Test.
4. Open the Permissions tab and configure user access. 👁 Updating permissions
   

Add a Personal Access Token

A Personal Access Token (PAT) authenticates MCP requests from Agno to CData Connect AI.

1. Open Settings and navigate to Access Tokens.
2. Click Create PAT.
3. Save the generated token securely. 👁 Creating a PAT
   

Step 2: Install dependencies and configure environment variables

Install Agno and the MCP adapter dependencies. LangChain is included strictly for MCP tool compatibility.

pip install agno agno-mcp langchain-mcp-adapters

Configure environment variables:

export CDATA_MCP_URL="https://mcp.cloud.cdata.com/mcp"
export CDATA_MCP_AUTH="Base64EncodedCredentials"
export OPENAI_API_KEY="your-openai-key"

Where "Base64EncodedCredentials" is your Connect AI user email and your Personal Access Token joined by a colon (":") and Base64 Encoded: Base64([email protected]:MY_CONNECT_AI_PAT)

Step 3: Connect to CData Connect AI via MCP

Create an MCP client using streamable HTTP. This establishes a secure connection to CData Connect AI.

import os
from langchain_mcp_adapters.client import MultiServerMCPClient

mcp_client = MultiServerMCPClient(
 connections={
 "default": {
 "transport": "streamable_http",
 "url": os.environ["CDATA_MCP_URL"],
 "headers": {
 "Authorization": f"Basic {os.environ['CDATA_MCP_AUTH']}"
 }
 }
 }
)

Step 4: Discover MCP tools

CData Connect AI exposes operations as MCP tools. These are retrieved dynamically at runtime.

langchain_tools = await mcp_client.get_tools()
for tool in langchain_tools:
 print(tool.name)

Step 5: Convert MCP tools to Agno functions

Each MCP tool is wrapped as an Agno function so it can be used by the agent.

NOTE: Agno performs all reasoning, planning, and tool selection.LangChain is used only as a lightweight MCP compatibility layer to consume tools exposed by CData Connect AI.

from agno.tools import Function

def make_tool_caller(lc_tool):
 async def call_tool(**kwargs):
 return await lc_tool.ainvoke(kwargs)
 return call_tool

Step 6: Create an Agno agent and query live Databricks data

Agno performs all reasoning, planning, and tool invocation. LangChain plays no role beyond MCP compatibility.

from agno.agent import Agent
from agno.models.openai import OpenAIChat

agent = Agent(
 model=OpenAIChat(
 id="gpt-4o",
 temperature=0.2,
 api_key=os.environ["OPENAI_API_KEY"]
 ),
 tools=agno_tools,
 markdown=True
)

await agent.aprint_response(
 "Show me the top 5 records from the available data source"
)

if __name__ == "__main__":
 asyncio.run(main())

The results below show an Agno agent invoking MCP tools through CData Connect AI and returning live Databricks data data.

👁 Running the Agno agent
👁 Agno agent output

You can now query live Databricks data using natural language through your Agno agent.

---

Get CData Connect AI

To get live data access to hundreds of SaaS, Big Data, and NoSQL sources directly from your cloud applications, try CData Connect AI today!