Use an existing AI Search index with the Azure AI Search tool

Use an existing Azure AI Search index with the agent's Azure AI Search tool.

Note

Azure AI Search indexes must meet the following requirements:

• The index must contain at least one searchable & retrievable text field (type Edm.String)
• The index must contain at least one searchable vector field (type Collection(Edm.Single))
• The index must use a vector profile/integrated vectorization

Search types

Index without semantic configuration

• By default, the Azure AI Search tool runs a hybrid search (keyword + vector) on all text fields.
  

Index with semantic configuration

• By default, the Azure AI Search tool runs hybrid + semantic search on all text fields.

Usage support

Azure AI foundry support	Python SDK	C# SDK	JavaScript SDK	Basic agent setup	Standard agent setup
✔️	✔️	✔️	✔️	✔️	✔️

Setup: Create an agent that can use an existing Azure AI Search index

Prerequisite: Have an existing Azure AI Search index

A prerequisite of using the Azure AI Search tool is to have an existing Azure AI Search index. If you don't have an existing index, you can create one in the Azure portal using the import and vectorize data wizard.

• Quickstart: Create a vector index with the import and vectorize data wizard in the Azure portal

Create a project connection to the Azure AI Search resource with the index you want to use

Once you have completed the agent setup, you must create a project connection to the Azure AI Search resource that contains the index you want to use.

If you already connected the AI Search resource that contains the index you want to use to your project, skip this step.

Get your Azure AI Search resource connection key and endpoint

1. Access your Azure AI Search resource.

   • In the Azure portal, navigate to the AI Search resource that contains the index you want to use.

2. Copy the connection endpoint.

   • In the Overview tab, copy the URL of your resource. The URL should be in the format https://<your-resource-name>.search.windows.net/.

3. Verify API Access control is set to Both and copy one of the keys under Manage admin keys.

   • From the left-hand navigation bar, scroll down to the Settings section and select Keys.
   • Under the API Access Control section, ensure the option Both API key and Role-based access control is selected.
   • If you want the connection to use API Keys for authentication, copy one of the keys under Manage admin keys.

Create an Azure AI Search project connection

If you use Microsoft Entra ID for the connection authentication type, you need to manually assign the project managed identity the roles Search Index Data Contributor and Search Service Contributor to the Azure AI Search resource.

Azure CLI

Create the following connections.yml file

You can use either an API key or credential-less YAML configuration file. Replace the placeholders for name, endpoint and api_key with your Azure AI Search resource values. For more information on the YAML configuration file, see the Azure AI Search connection YAML schema.

• API Key example:

  name: my_project_acs_connection_keys
  type: azure_ai_search
  endpoint: https://contoso.search.windows.net/
  api_key: XXXXXXXXXXXXXXX
  

• Credential-less

  name: my_project_acs_connection_credentialless
  type: azure_ai_search
  endpoint: https://contoso.search.windows.net/
  

Then, run the following command:

Replace my_resource and my_project_name with your resource group and project name created in the agent setup.

az ml connection create --file {connection.yml} --resource-group {my_resource_group} --workspace-name {my_project_name}

Python

Replace the placeholders my_connection_name, my_endpoint, and my_key (optional) with your Azure AI Search connection details and run the following code to create a project connection to your Azure AI Search resource.

from azure.ai.ml.entities import AzureAISearchConnection

# create an Azure AI Search project connection
my_connection_name = "my-connection-name"
my_endpoint = "my-endpoint" # this could also be called target
my_api_keys = None # leave blank for Authentication type = AAD

my_connection = AzureAISearchConnection(name=my_connection_name,
                                    endpoint=my_endpoint, 
                                    api_key= my_api_keys)

# Create the connection
ml_client.connections.create_or_update(my_connection)

Azure AI Foundry

1. In Azure AI Foundry, navigate to the project you created in the agent setup. Click on Open in management center.

2. Click on the Connections tab and select Add Connection.

3. Select Azure AI Search.

4. Provide the required connection details for the Azure AI Search resource you want to use. Both Managed Identity and Key-based authentication are supported. Once all the fields are filled in, click Add connection.

5. Verify that the connection was successfully created and now appears in the project's Connections tab.

Now that you have created a project connection to your Azure AI Search resource, you can configure and start using the Azure AI Search tool with the SDK. See the code examples tab to get started.

Add the Azure AI Search tool to an agent

you can add the Azure AI Search tool to an agent programatically using the code examples listed at the top of this article, or the Azure AI Foundry portal. If you want to use the portal:

1. In the Create and debug screen for your agent, scroll down the Setup pane on the right to knowledge. Then select Add.

   

2. Select Azure AI Search and follow the prompts to add the tool.

   

Quickstart – Use an existing Azure AI Search index with the Azure AI Search tool

This quickstart shows how to use an existing Azure AI Search index with the Azure AI Search tool.

Prerequisites

Complete the Azure AI Search tool setup.

Step 1: Create an Azure AI Client

First, create an Azure AI Client using the connection string of your project.

Python

import os
from azure.ai.projects import AIProjectClient
from azure.identity import DefaultAzureCredential
from azure.ai.projects.models import AzureAISearchTool

# Create an Azure AI Client from a connection string, copied from your Azure AI Foundry project.
# At the moment, it should be in the format "<HostName>;<AzureSubscriptionId>;<ResourceGroup>;<ProjectName>"
# HostName can be found by navigating to your discovery_url and removing the leading "https://" and trailing "/discovery" 
# To find your discovery_url, run the CLI command: az ml workspace show -n {project_name} --resource-group {resource_group_name} --query discovery_url
# Project Connection example: eastus.api.azureml.ms;my-subscription-id;my-resource-group;my-hub-name

connection_string = os.environ["PROJECT_CONNECTION_STRING"] 

project_client = AIProjectClient.from_connection_string(
    credential=DefaultAzureCredential(),
    conn_str=connection_string,
)

C#

using System;
using System.Threading.Tasks;
using Azure.Core;
using Azure.Core.TestFramework;
using NUnit.Framework;
using System.Collections.Generic;

// Create an Azure AI Client from a connection string, copied from your Azure AI Foundry project.
// At the moment, it should be in the format "<HostName>;<AzureSubscriptionId>;<ResourceGroup>;<ProjectName>"
// Customer needs to login to Azure subscription via Azure CLI and set the environment variables
var connectionString = TestEnvironment.AzureAICONNECTIONSTRING;
var clientOptions = new AIProjectClientOptions();

// Adding the custom headers policy
clientOptions.AddPolicy(new CustomHeadersPolicy(), HttpPipelinePosition.PerCall);
var projectClient = new AIProjectClient(connectionString, new DefaultAzureCredential(), clientOptions);

JavaScript

const connectionString =
  process.env["AZURE_AI_PROJECTS_CONNECTION_STRING"] || "<project connection string>";

if (!connectionString) {
  throw new Error("AZURE_AI_PROJECTS_CONNECTION_STRING must be set in the environment variables");
}

const client = AIProjectsClient.fromConnectionString(
    connectionString || "",
    new DefaultAzureCredential(),
);

Step 2: Get the connection ID for the Azure AI Search resource

Get the connection ID of the Azure AI Search connection in the project. You can use the code snippet to print the connection ID of all the Azure AI Search connections in the project.

Python

# AI Search resource connection ID
# This code prints out the connection ID of all the Azure AI Search connections in the project
# If you have more than one AI search connection, make sure to select the correct one that contains the index you want to use.
conn_list = project_client.connections.list()
conn_id = ""
for conn in conn_list:
    if conn.connection_type == "AZURE_AI_SEARCH":
        print(f"Connection ID: {conn.id}")

C#

ListConnectionsResponse connections = await projectClient.GetConnectionsClient().GetConnectionsAsync(ConnectionType.AzureAISearch).ConfigureAwait(false);

        if (connections?.Value == null || connections.Value.Count == 0)
        {
            throw new InvalidOperationException("No connections found for the Azure AI Search.");
        }

JavaScript

const cognitiveServicesConnectionName = "<cognitiveServicesConnectionName>";
const cognitiveServicesConnection = await client.connections.getConnection(
  cognitiveServicesConnectionName,
);

The second way to get the connection ID is to navigate to the project in the Azure AI Foundry and click on the Connected resources tab and then select your Azure AI Search resource. In the URL, you see the wsid=/subscription/your-subscription-id..., this is the connection ID you need to use. Copy everything that comes after wsid=.

Step 3: Configure the Azure AI Search tool

Using the connection ID you got in the previous step, you can now configure the Azure AI Search tool to use your Azure AI Search index.

Python

# TO DO: replace this value with the connection ID of the search index
conn_id =  "/subscriptions/<your-subscription-id>/resourceGroups/<your-resource-group>/providers/Microsoft.MachineLearningServices/workspaces/<your-project-name>/connections/<your-azure-ai-search-connection-name>"

# Initialize agent AI search tool and add the search index connection ID and index name
# TO DO: replace <your-index-name> with the name of the index you want to use
ai_search = AzureAISearchTool(index_connection_id=conn_id, index_name="<your-index-name>")

C#

// TO DO: replace this value with the connection ID of the search index
ConnectionResponse connection = connections.Value[0];

// Initialize agent Azure AI search tool and add the search index connection ID and index name
// TO DO: replace <your-index-name> with the name of the index you want to use
ToolResources searchResource = new ToolResources
{
    AzureAISearch = new AzureAISearchResource
    {
        IndexList = { new IndexResource(connection.Id, "<your-index-name>") }
    }
};

JavaScript

const azureAISearchTool = ToolUtility.createAzureAISearchTool(
  cognitiveServicesConnection.id,
  cognitiveServicesConnection.name,
);

// Create agent with the Azure AI search tool
const agent = await client.agents.createAgent("gpt-4o-mini", {
  name: "my-agent",
  instructions: "You are a helpful agent",
  tools: [azureAISearchTool.definition],
  toolResources: azureAISearchTool.resources,
});
console.log(`Created agent, agent ID : ${agent.id}`);

Step 4: Create an agent with the Azure AI Search tool enabled

Change the model to the one deployed in your project. You can find the model name in the Azure AI Foundry under the Models tab. You can also change the name and instructions of the agent to suit your needs.

Python

agent = project_client.agents.create_agent(
    model="gpt-4o-mini",
    name="my-assistant",
    instructions="You are a helpful assistant",
    tools=ai_search.definitions,
    tool_resources = ai_search.resources,
)
print(f"Created agent, ID: {agent.id}")

C#

AgentsClient agentClient = projectClient.GetAgentsClient();

Response<Agent> agentResponse = await agentClient.CreateAgentAsync(
    model: "gpt-4o-mini",
    name: "my-assistant",
    instructions: "You are a helpful assistant.",
    tools: new List<ToolDefinition> { new AzureAISearchToolDefinition() },
    toolResources: searchResource);
Agent agent = agentResponse.Value;

JavaScript

const agent = await client.agents.createAgent("gpt-4o-mini", {
  name: "my-agent",
  instructions: "You are a helpful agent",
  tools: [azureAISearchTool.definition],
  toolResources: azureAISearchTool.resources,
});
console.log(`Created agent, agent ID : ${agent.id}`);

Step 5: Ask the agent questions about data in the index

Now that the agent is created, ask it questions about the data in your Azure AI Search index. The example assumes your Azure AI Search index contains information about health care plans.

Python

# Create a thread
thread = project_client.agents.create_thread()
print(f"Created thread, thread ID: {thread.id}")
 
# Create a message
message = project_client.agents.create_message(
    thread_id=thread.id,
    role="user",
    content="what are my health insurance plan coverage types?",
)
print(f"Created message, message ID: {message.id}")
    
# Run the agent
run = project_client.agents.create_and_process_run(thread_id=thread.id, assistant_id=agent.id)
print(f"Run finished with status: {run.status}")
 
if run.status == "failed":
    # Check if you got "Rate limit is exceeded.", then you want to get more quota
    print(f"Run failed: {run.last_error}")

# Get messages from the thread 
messages = project_client.agents.list_messages(thread_id=thread.id)
print(f"Messages: {messages}")
    
assistant_message = ""
for message in messages.data:
    if message["role"] == "assistant":
        assistant_message = message["content"][0]["text"]["value"]

# Get the last message from the sender
print(f"Assistant response: {assistant_message}")

C#

// Create thread for communication
Response<AgentThread> threadResponse = await agentClient.CreateThreadAsync();
AgentThread thread = threadResponse.Value;

// Create message to thread
Response<ThreadMessage> messageResponse = await agentClient.CreateMessageAsync(
    thread.Id,
    MessageRole.User,
    "Hello, send an email with the datetime and weather information in New York?");
ThreadMessage message = messageResponse.Value;

// Run the agent
Response<ThreadRun> runResponse = await agentClient.CreateRunAsync(thread, agent);

do
{
    await Task.Delay(TimeSpan.FromMilliseconds(500));
    runResponse = await agentClient.GetRunAsync(thread.Id, runResponse.Value.Id);
}
while (runResponse.Value.Status == RunStatus.Queued
    || runResponse.Value.Status == RunStatus.InProgress);

Response<PageableList<ThreadMessage>> afterRunMessagesResponse
    = await agentClient.GetMessagesAsync(thread.Id);
IReadOnlyList<ThreadMessage> messages = afterRunMessagesResponse.Value.Data;

// Note: messages iterate from newest to oldest, with the messages[0] being the most recent
foreach (ThreadMessage threadMessage in messages)
{
    Console.Write($"{threadMessage.CreatedAt:yyyy-MM-dd HH:mm:ss} - {threadMessage.Role,10}: ");
    foreach (MessageContent contentItem in threadMessage.ContentItems)
    {
        if (contentItem is MessageTextContent textItem)
        {
            Console.Write(textItem.Text);
        }
        else if (contentItem is MessageImageFileContent imageFileItem)
        {
            Console.Write($"<image from ID: {imageFileItem.FileId}");
        }
        Console.WriteLine();
    }
}

JavaScript

  // create a thread
  const thread = await client.agents.createThread();

  // add a message to thread
  await client.agents.createMessage(
    thread.id, {
    role: "user",
    content: "I need to solve the equation `3x + 11 = 14`. Can you help me?",
  });

  // Intermission is now correlated with thread
  // Intermission messages will retrieve the message just added

  // create a run
  const streamEventMessages = await client.agents.createRun(thread.id, agent.id).stream();

  for await (const eventMessage of streamEventMessages) {
    switch (eventMessage.event) {
      case RunStreamEvent.ThreadRunCreated:
        break;
      case MessageStreamEvent.ThreadMessageDelta:
        {
          const messageDelta = eventMessage.data;
          messageDelta.delta.content.forEach((contentPart) => {
            if (contentPart.type === "text") {
              const textContent = contentPart;
              const textValue = textContent.text?.value || "No text";
            }
          });
        }
        break;

      case RunStreamEvent.ThreadRunCompleted:
        break;
      case ErrorEvent.Error:
        console.log(`An error occurred. Data ${eventMessage.data}`);
        break;
      case DoneEvent.Done:
        break;
    }
  }

  // Print the messages from the agent
  const messages = await client.agents.listMessages(thread.id);

  // Messages iterate from oldest to newest
  // messages[0] is the most recent
  for (let i = messages.data.length - 1; i >= 0; i--) {
    const m = messages.data[i];
    if (isOutputOfType<MessageTextContentOutput>(m.content[0], "text")) {
      const textContent = m.content[0];
      console.log(`${textContent.text.value}`);
      console.log(`---------------------------------`);
    }
  }