Guida introduttiva: Creare un agente di prompt

In questa guida rapida si crea un agente prompt nel servizio Foundry Agent e si avvia una conversazione con esso. Un agente prompt è un agente definito in modo dichiarativo che combina un modello dal catalogo dei modelli di Foundry, con istruzioni, strumenti e prompt di linguaggio naturale per orientare il comportamento.

Se non si ha una sottoscrizione Azure, creare un account free.

Prerequisiti

Modello distribuito in Microsoft Foundry. Se non si dispone di un modello, completare prima di tutto Guida rapida: Configurare le risorse di Microsoft Foundry.
I runtime del linguaggio necessari, gli strumenti globali e le estensioni Visual Studio Code come descritto in Prepare l'ambiente di sviluppo.

Impostare le variabili di ambiente

Archiviare l'endpoint del progetto come variabile di ambiente. Impostare anche questi valori per l'uso negli script.

Python e JavaScript

PROJECT_ENDPOINT=<endpoint copied from welcome screen>
AGENT_NAME="MyAgent"

C# e Java

ProjectEndpoint = <endpoint copied from welcome screen>
AgentName = "MyAgent"

Installare pacchetti ed eseguire l'autenticazione

Assicurarsi di installare la versione corretta dei pacchetti, come illustrato di seguito.

Installare la versione corrente di azure-ai-projects. Questa versione usa l'API Foundry projects (new) .
```
pip install azure-ai-projects>=2.0.0
```
Accedi utilizzando il comando CLI az login per autenticarti prima di eseguire i tuoi script Python.

Installare i pacchetti:

Aggiungere pacchetti NuGet usando l'interfaccia della riga di comando di .NET nel terminale integrato: questi pacchetti usano l'API Foundry projects (new) API.
```
dotnet add package Azure.AI.Projects
dotnet add package Azure.AI.Projects.Agents
dotnet add package Azure.AI.Extensions.OpenAI
dotnet add package Azure.Identity
```
Accedi utilizzando il comando CLI az login per autenticarti prima di lanciare gli script C#.

Installare la versione corrente di @azure/ai-projects. Questa versione usa l'API Dei progetti Foundry (nuova).
```
npm install @azure/ai-projects
```
Accedere usando il comando CLI az login per eseguire l'autenticazione prima di eseguire gli script TypeScript.

<dependency>
    <groupId>com.azure</groupId>
    <artifactId>azure-ai-agents</artifactId>
    <version>2.0.0</version>
</dependency>

Accedi usando il comando CLI az login per autenticarti prima di eseguire gli script Java.

Accedi usando il comando CLI az login per autenticarti prima di eseguire il comando successivo.
Ottenere un token di accesso temporaneo. Scadrà tra 60 e 90 minuti, sarà necessario eseguire l'aggiornamento dopo.
```
az account get-access-token --scope https://ai.azure.com/.default
```
Salvare i risultati come variabile AZURE_AI_AUTH_TOKENdi ambiente .

Suggerimento

Il codice usa Azure Ai Projects 2.x ed è incompatibile con Azure AI Projects 1.x. Visualizza la documentazione di Foundry (versione classica) per la versione Azure AI Projects 1.x.

Creare un agente di risposta rapida

Crea un agente per i prompt utilizzando il tuo modello distribuito. L'agente usa un PromptAgentDefinition oggetto con istruzioni che definiscono il comportamento dell'agente. È possibile aggiornare o eliminare agenti in qualsiasi momento.

from azure.identity import DefaultAzureCredential
from azure.ai.projects import AIProjectClient
from azure.ai.projects.models import PromptAgentDefinition

# Format: "https://resource_name.ai.azure.com/api/projects/project_name"
PROJECT_ENDPOINT = "your_project_endpoint"
AGENT_NAME = "your_agent_name"

# Create project client to call Foundry API
project = AIProjectClient(
    endpoint=PROJECT_ENDPOINT,
    credential=DefaultAzureCredential(),
)

# Create an agent with a model and instructions
agent = project.agents.create_version(
    agent_name=AGENT_NAME,
    definition=PromptAgentDefinition(
        model="gpt-5-mini",  # supports all Foundry direct models"
        instructions="You are a helpful assistant that answers general questions",
    ),
)
print(f"Agent created (id: {agent.id}, name: {agent.name}, version: {agent.version})")

using Azure.Identity;
using Azure.AI.Projects;
using Azure.AI.Projects.Agents;
using Azure.AI.Extensions.OpenAI;

// Format: "https://resource_name.ai.azure.com/api/projects/project_name"
var ProjectEndpoint = "your_project_endpoint";
var AgentName = "your_agent_name";

// Create project client to call Foundry API
AIProjectClient projectClient = new(
    endpoint: new Uri(ProjectEndpoint),
    tokenProvider: new DefaultAzureCredential());

// Create an agent with a model and instructions
ProjectsAgentDefinition agentDefinition = new DeclarativeAgentDefinition("gpt-5-mini") // supports all Foundry direct models
{
    Instructions = "You are a helpful assistant that answers general questions",
};

ProjectsAgentVersion agent = projectClient.AgentAdministrationClient.CreateAgentVersion(
    AgentName,
    options: new(agentDefinition));
Console.WriteLine($"Agent created (id: {agent.Id}, name: {agent.Name}, version: {agent.Version})");

import { DefaultAzureCredential } from "@azure/identity";
import { AIProjectClient } from "@azure/ai-projects";

// Format: "https://resource_name.ai.azure.com/api/projects/project_name"
const PROJECT_ENDPOINT = "your_project_endpoint";
const AGENT_NAME = "your_agent_name";

async function main(): Promise<void> {
    // Create project client to call Foundry API
    const project = new AIProjectClient(PROJECT_ENDPOINT, new DefaultAzureCredential());

    // Create an agent with a model and instructions
    const agent = await project.agents.createVersion(AGENT_NAME, {
        kind: "prompt",
        model: "gpt-5-mini", //supports all Foundry direct models
        instructions: "You are a helpful assistant that answers general questions",
    });
    console.log(`Agent created (id: ${agent.id}, name: ${agent.name}, version: ${agent.version})`);
}

main().catch(console.error);

package com.azure.ai.agents;

import com.azure.ai.agents.models.AgentVersionDetails;
import com.azure.ai.agents.models.PromptAgentDefinition;
import com.azure.identity.DefaultAzureCredentialBuilder;

public class CreateAgent {
    public static void main(String[] args) {
        // Format: "https://resource_name.ai.azure.com/api/projects/project_name"
        String ProjectEndpoint = "your_project_endpoint";
        String AgentName = "your_agent_name";

        // Create agents client to call Foundry API
        AgentsClient agentsClient = new AgentsClientBuilder()
                .credential(new DefaultAzureCredentialBuilder().build())
                .endpoint(ProjectEndpoint)
                .buildAgentsClient();

        // Create an agent with a model and instructions
        PromptAgentDefinition request = new PromptAgentDefinition("gpt-5-mini") // supports all Foundry direct models
                .setInstructions("You are a helpful assistant that answers general questions");
        AgentVersionDetails agent = agentsClient.createAgentVersion(AgentName, request);

        System.out.println("Agent ID: " + agent.getId());
        System.out.println("Agent Name: " + agent.getName());
        System.out.println("Agent Version: " + agent.getVersion());
    }
}

Sostituire YOUR-FOUNDRY-RESOURCE-NAME con i valori:

curl -X POST https://YOUR-FOUNDRY-RESOURCE-NAME.services.ai.azure.com/api/projects/YOUR-PROJECT-NAME/agents?api-version=v1 \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AZURE_AI_AUTH_TOKEN" \
  -d '{
    "name": "MyAgent",
    "definition": {
      "kind": "prompt",
      "model": "gpt-4.1-mini",
      "instructions": "You are a helpful assistant that answers general questions"
    }
  }'

L'output conferma che l'agente è stato creato. Nella console viene visualizzato il nome e l'ID dell'agente.

Chat con l'agente

Usa l'agente che hai creato per interagire ponendo una domanda e un'interazione successiva correlata. La conversazione mantiene la cronologia di queste interazioni.

from azure.identity import DefaultAzureCredential
from azure.ai.projects import AIProjectClient

# Format: "https://resource_name.ai.azure.com/api/projects/project_name"
PROJECT_ENDPOINT = "your_project_endpoint"
AGENT_NAME = "your_agent_name"

# Create project and openai clients to call Foundry API
project = AIProjectClient(
    endpoint=PROJECT_ENDPOINT,
    credential=DefaultAzureCredential(),
)
openai = project.get_openai_client()

# Create a conversation for multi-turn chat
conversation = openai.conversations.create()

# Chat with the agent to answer questions
response = openai.responses.create(
    conversation=conversation.id,
    extra_body={"agent_reference": {"name": AGENT_NAME, "type": "agent_reference"}},
    input="What is the size of France in square miles?",
)
print(response.output_text)

# Ask a follow-up question in the same conversation
response = openai.responses.create(
    conversation=conversation.id,
    extra_body={"agent_reference": {"name": AGENT_NAME, "type": "agent_reference"}},
    input="And what is the capital city?",
)
print(response.output_text)

using Azure.Identity;
using Azure.AI.Projects;
using Azure.AI.Extensions.OpenAI;
using OpenAI.Responses;

#pragma warning disable OPENAI001

// Format: "https://resource_name.ai.azure.com/api/projects/project_name"
var ProjectEndpoint = "your_project_endpoint";
var AgentName = "your_agent_name";

// Create project client to call Foundry API
AIProjectClient projectClient = new(
    endpoint: new Uri(ProjectEndpoint),
    tokenProvider: new DefaultAzureCredential());

// Create a conversation for multi-turn chat
ProjectConversation conversation = projectClient.ProjectOpenAIClient.GetProjectConversationsClient().CreateProjectConversation();

// Chat with the agent to answer questions
ProjectResponsesClient responsesClient = projectClient.ProjectOpenAIClient.GetProjectResponsesClientForAgent(
    defaultAgent: AgentName,
    defaultConversationId: conversation.Id);
ResponseResult response = responsesClient.CreateResponse("What is the size of France in square miles?");
Console.WriteLine(response.GetOutputText());

// Ask a follow-up question in the same conversation
response = responsesClient.CreateResponse("And what is the capital city?");
Console.WriteLine(response.GetOutputText());

import { DefaultAzureCredential } from "@azure/identity";
import { AIProjectClient } from "@azure/ai-projects";

// Format: "https://resource_name.ai.azure.com/api/projects/project_name"
const PROJECT_ENDPOINT = "your_project_endpoint";
const AGENT_NAME = "your_agent_name";

async function main(): Promise<void> {
    // Create project and openai clients to call Foundry API
    const project = new AIProjectClient(PROJECT_ENDPOINT, new DefaultAzureCredential());
    const openai = project.getOpenAIClient();

    // Create a conversation for multi-turn chat
    const conversation = await openai.conversations.create();

    // Chat with the agent to answer questions
    const response = await openai.responses.create(
        {
            conversation: conversation.id,
            input: "What is the size of France in square miles?",
        },
        {
            body: { agent_reference: { name: AGENT_NAME, type: "agent_reference" } },
        },
    );
    console.log(response.output_text);

    // Ask a follow-up question in the same conversation
    const response2 = await openai.responses.create(
        {
            conversation: conversation.id,
            input: "And what is the capital city?",
        },
        {
            body: { agent_reference: { name: AGENT_NAME, type: "agent_reference" } },
        },
    );
    console.log(response2.output_text);
}

main().catch(console.error);

package com.azure.ai.agents;

import com.azure.ai.agents.models.AgentDetails;
import com.azure.ai.agents.models.AgentReference;
import com.azure.ai.agents.models.AzureCreateResponseOptions;
import com.azure.ai.agents.models.AgentVersionDetails;
import com.azure.ai.agents.models.PromptAgentDefinition;
import com.azure.identity.AuthenticationUtil;
import com.azure.identity.DefaultAzureCredentialBuilder;
import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.credential.BearerTokenCredential;
import com.openai.models.conversations.Conversation;
import com.openai.models.conversations.items.ItemCreateParams;
import com.openai.models.responses.EasyInputMessage;
import com.openai.models.responses.Response;
import com.openai.models.responses.ResponseCreateParams;
import com.openai.services.blocking.ConversationService;

public class ChatWithAgent {
    public static void main(String[] args) {
        // Format: "https://resource_name.ai.azure.com/api/projects/project_name"
        String ProjectEndpoint = "your_project_endpoint";
        String AgentName = "your_agent_name";
        
        AgentsClientBuilder builder = new AgentsClientBuilder()
                .credential(new DefaultAzureCredentialBuilder().build())
                .endpoint(ProjectEndpoint);

        AgentsClient agentsClient = builder.buildAgentsClient();
        ResponsesClient responsesClient = builder.buildResponsesClient();
        ConversationService conversationService
            = builder.buildOpenAIClient().conversations();

        AgentDetails agent = agentsClient.getAgent(AgentName);

        Conversation conversation = conversationService.create();
        conversationService.items().create(
            ItemCreateParams.builder()
                .conversationId(conversation.id())
                .addItem(EasyInputMessage.builder()
                    .role(EasyInputMessage.Role.SYSTEM)
                    .content("You are a helpful assistant that speaks like a pirate.")
                    .build()
                ).addItem(EasyInputMessage.builder()
                    .role(EasyInputMessage.Role.USER)
                    .content("Hello, agent!")
                    .build()
            ).build()
        );

        AgentReference agentReference = new AgentReference(agent.getName()).setVersion(agent.getVersion());
        Response response = responsesClient.createAzureResponse(
            new AzureCreateResponseOptions().setAgentReference(agentReference),
            ResponseCreateParams.builder().conversation(conversation.id()));

        OpenAIClient client = OpenAIOkHttpClient.builder()
            .baseUrl(ProjectEndpoint.endsWith("/") ? ProjectEndpoint + "openai/v1" : ProjectEndpoint + "/openai/v1")
            .credential(BearerTokenCredential.create(AuthenticationUtil.getBearerTokenSupplier(
                    new DefaultAzureCredentialBuilder().build(), "https://ai.azure.com/.default")))
            .build();

        ResponseCreateParams responseRequest = new ResponseCreateParams.Builder()
            .input("Hello, how can you help me?")
            .model("gpt-5-mini") //supports all Foundry direct models
            .build();

        Response result = client.responses().create(responseRequest);
    }
}

Sostituire YOUR-FOUNDRY-RESOURCE-NAME con i valori:

# Generate a response using the agent

curl -X POST https://YOUR-FOUNDRY-RESOURCE-NAME.services.ai.azure.com/api/projects/YOUR-PROJECT-NAME/openai/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AZURE_AI_AUTH_TOKEN" \
  -d '{
    "agent_reference": {"type": "agent_reference", "name": "<AGENT_NAME>"},
    "input": [{"role": "user", "content": "What is the size of France in square miles?"}]
  }'

# Optional Step: Create a conversation to use with the agent
curl -X POST https://YOUR-FOUNDRY-RESOURCE-NAME.services.ai.azure.com/api/projects/YOUR-PROJECT-NAME/openai/v1/conversations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AZURE_AI_AUTH_TOKEN" \
  -d '{
    "items": [
      {
        "type": "message",
        "role": "user",
        "content": [
          {
            "type": "input_text",
            "text": "What is the size of France in square miles?"
          }
        ]
      }
    ]
  }'

# Lets say Conversation ID created is conv_123456789. Use this in the next step

#Optional Step: Ask a follow-up question in the same conversation
curl -X POST https://YOUR-FOUNDRY-RESOURCE-NAME.services.ai.azure.com/api/projects/YOUR-PROJECT-NAME/openai/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AZURE_AI_AUTH_TOKEN" \
  -d '{
    "agent_reference": {"type": "agent_reference", "name": "<AGENT_NAME>", "version": "1"},
    "conversation": "<CONVERSATION_ID>",
    "input": [{"role": "user", "content": "And what is the capital?"}]
  }'

Vengono visualizzate le risposte dell'agente a entrambe le richieste. La risposta di follow-up dimostra che l'agente mantiene la cronologia delle conversazioni tra i turni.

Pulire le risorse

Se non sono più necessarie risorse create, eliminare il gruppo di risorse associato al progetto.

Nel portale Azure portale selezionare il gruppo di risorse e quindi selezionare Delete. Verificare di voler eliminare il gruppo di risorse.

Commenti e suggerimenti

Questa pagina è stata utile?

Last updated on 2026-05-01