Skip to main content

​
Azure Functions Tool

Azure Functions enables serverless compute for agent tools, providing enterprise integration capabilities with custom code execution.

​
When to Use Azure Functions

Choose Azure Functions over function calling when you need:
  • Separation of concerns: Isolate business logic from agent code
  • Centralized management: Share tools across multiple agents
  • Security isolation: Control agent and tool permissions separately
  • External dependencies: Use non-Microsoft libraries or legacy systems
  • Complex operations: Handle multi-step workflows
  • Async processing: Execute long-running operations

​
Integration Options

​
Queue-Based Tools

Asynchronous processing via Azure Queue Storage: 
from azure.ai.projects.models import AzureFunctionTool

function_tool = AzureFunctionTool(
    name="process_order",
    description="Process a customer order",
    input_binding={"storage_queue": "order-queue"},
    output_binding={"storage_queue": "order-results"}
)
Best for: Background processing, reliable message delivery, retry capabilities

​
HTTP-Based Tools

Synchronous calls via HTTP triggers (use OpenAPI tool) Best for: Real-time responses, immediate feedback

​
MCP Server

Model Context Protocol for reusable tool catalogs Best for: Industry standard, shared tools across agents

​
Quick Start Template

1

Initialize Template

azd init --template azure-functions-ai-services-agent-python
2

Provision Resources

azd provision
Creates: Function app, Foundry resources, Storage, Cosmos DB, AI Search
3

Deploy Code

azd deploy

​
Prerequisites

  • Azure Functions Core Tools v4.x
  • Deployed agent with Standard Setup (Basic not supported)
  • Azure account with active subscription
See Function Calling for in-process functions.

Build docs developers (and LLMs) love

Get started for freeTalk to us