Del via

Python code interpreter agent tool

Azure Databricks provides system.ai.python_exec, a built-in Unity Catalog function that lets AI agents dynamically run Python code written by the agent, provided by a user, or retrieved from a codebase. It is available by default and can be used directly in a SQL query:

SELECT python_exec("""
import random
numbers = [random.random() for _ in range(10)]
print(numbers)
""")

To learn more about agent tools, see AI agent tools.

Add the code interpreter to your agent

To add python_exec to your agent, connect to the managed MCP server for the system.ai Unity Catalog schema. The code interpreter is available as a pre-configured MCP tool at https://<workspace-hostname>/api/2.0/mcp/functions/system/ai/python_exec.

OpenAI Agents SDK (Apps)

from agents import Agent, Runner
from databricks.sdk import WorkspaceClient
from databricks_openai.agents import McpServer

# WorkspaceClient picks up credentials from the environment (Databricks Apps, notebook, CLI)
workspace_client = WorkspaceClient()
host = workspace_client.config.host

# The context manager manages the MCP connection lifecycle and ensures cleanup on exit.
# from_uc_function constructs the endpoint URL from UC identifiers and wires in auth
# from workspace_client, avoiding hardcoded URLs and manual token handling.
async with McpServer.from_uc_function(
    catalog="system",
    schema="ai",
    function_name="python_exec",
    workspace_client=workspace_client,
    name="code-interpreter",
) as code_interpreter:
    agent = Agent(
        name="Coding agent",
        instructions="You are a helpful coding assistant. Use the python_exec tool to run code.",
        model="databricks-claude-sonnet-4-5",
        mcp_servers=[code_interpreter],
    )
    result = await Runner.run(agent, "Calculate the first 10 Fibonacci numbers")
    print(result.final_output)

Grant the app access to the function in databricks.yml:

resources:
  apps:
    my_agent_app:
      resources:
        - name: 'python_exec'
          uc_securable:
            securable_full_name: 'system.ai.python_exec'
            securable_type: 'FUNCTION'
            permission: 'EXECUTE'

LangGraph (Apps)

from databricks.sdk import WorkspaceClient
from databricks_langchain import ChatDatabricks, DatabricksMCPServer, DatabricksMultiServerMCPClient
from langgraph.prebuilt import create_react_agent

workspace_client = WorkspaceClient()
host = workspace_client.config.host

# DatabricksMultiServerMCPClient provides a unified get_tools() interface across
# multiple MCP servers, making it easy to add more servers later without refactoring.
mcp_client = DatabricksMultiServerMCPClient([
    DatabricksMCPServer(
        name="code-interpreter",
        url=f"{host}/api/2.0/mcp/functions/system/ai/python_exec",
        workspace_client=workspace_client,
    ),
])

async with mcp_client:
    tools = await mcp_client.get_tools()
    agent = create_react_agent(
        ChatDatabricks(endpoint="databricks-claude-sonnet-4-5"),
        tools=tools,
    )
    result = await agent.ainvoke(
        {"messages": [{"role": "user", "content": "Calculate the first 10 Fibonacci numbers"}]}
    )
    # LangGraph returns the full conversation history; the last message is the agent's final response
    print(result["messages"][-1].content)

Grant the app access to the function in databricks.yml:

resources:
  apps:
    my_agent_app:
      resources:
        - name: 'python_exec'
          uc_securable:
            securable_full_name: 'system.ai.python_exec'
            securable_type: 'FUNCTION'
            permission: 'EXECUTE'

Model Serving

from databricks.sdk import WorkspaceClient
from databricks_mcp import DatabricksMCPClient
import mlflow

workspace_client = WorkspaceClient()
host = workspace_client.config.host

mcp_client = DatabricksMCPClient(
    server_url=f"{host}/api/2.0/mcp/functions/system/ai/python_exec",
    workspace_client=workspace_client,
)

tools = mcp_client.list_tools()

# get_databricks_resources() extracts the UC permissions the agent needs at runtime.
# Passing these to log_model lets Model Serving grant access automatically at deployment,
# without requiring manual permission configuration.
mlflow.pyfunc.log_model(
    "agent",
    python_model=my_agent,
    resources=mcp_client.get_databricks_resources(),
)

To deploy the agent, see Deploy an agent for generative AI applications (Model Serving). For details on logging agents with MCP resources, see Use Databricks managed MCP servers.

Next steps

  • Last updated on