通过


OpenAI 代理

Microsoft Agent Framework 支持三种不同的 OpenAI 客户端类型,每个类型面向具有不同工具功能的不同 API 图面:

客户端类型 API 最适用于
聊天完成 聊天完成 API 简单代理,广泛的模型支持
反应 响应 API 具有托管工具的全功能代理(代码解释器、文件搜索、Web 搜索、托管 MCP)
助手 助手 API 具有代码解释器和文件搜索的服务器托管代理

小窍门

有关 Azure OpenAI 的等效项(AzureOpenAIChatClientAzureOpenAIResponsesClientAzureOpenAIAssistantsClient),可以查看 Azure OpenAI 提供程序页以获取更多信息。 工具支持是相同的。

入门

将所需的 NuGet 包添加到项目。

dotnet add package Microsoft.Agents.AI.OpenAI --prerelease

聊天完成客户端

聊天完成客户端提供了使用 ChatCompletion API 创建代理的简单方法。

using Microsoft.Agents.AI;
using OpenAI;

OpenAIClient client = new OpenAIClient("<your_api_key>");
var chatClient = client.GetChatClient("gpt-4o-mini");

AIAgent agent = chatClient.AsAIAgent(
    instructions: "You are good at telling jokes.",
    name: "Joker");

Console.WriteLine(await agent.RunAsync("Tell me a joke about a pirate."));

支持的工具: 函数工具、Web 搜索、本地 MCP 工具。

响应客户端

响应客户端提供最丰富的工具支持,包括代码解释器、文件搜索、Web 搜索和托管 MCP。

using Microsoft.Agents.AI;
using OpenAI;

OpenAIClient client = new OpenAIClient("<your_api_key>");
var responsesClient = client.GetResponseClient("gpt-4o-mini");

AIAgent agent = responsesClient.AsAIAgent(
    instructions: "You are a helpful coding assistant.",
    name: "CodeHelper");

Console.WriteLine(await agent.RunAsync("Write a Python function to sort a list."));

支持的工具: 函数工具、工具审批、代码解释器、文件搜索、Web 搜索、托管 MCP、本地 MCP 工具。

助手客户端

Assistants 客户端使用内置代码解释器和文件搜索创建服务器管理的代理。

using Microsoft.Agents.AI;
using OpenAI;

OpenAIClient client = new OpenAIClient("<your_api_key>");
var assistantsClient = client.GetAssistantClient();

// Assistants are managed server-side
AIAgent agent = assistantsClient.AsAIAgent(
    instructions: "You are a data analysis assistant.",
    name: "DataHelper");

Console.WriteLine(await agent.RunAsync("Analyze trends in the uploaded data."));

支持的工具: 函数工具、代码解释器、文件搜索、本地 MCP 工具。

小窍门

有关完整的可运行示例,请参阅 .NET 示例

使用代理

这三种客户端类型都生成一个支持相同代理操作(流式处理、线程、中间件)的标准 AIAgent

有关详细信息,请参阅 入门教程

安装

pip install agent-framework --pre

配置

每个客户端类型使用不同的环境变量:

聊天补全

OPENAI_API_KEY="your-openai-api-key"
OPENAI_CHAT_MODEL_ID="gpt-4o-mini"

Responses

OPENAI_API_KEY="your-openai-api-key"
OPENAI_RESPONSES_MODEL_ID="gpt-4o-mini"

Assistants

OPENAI_API_KEY="your-openai-api-key"
OPENAI_CHAT_MODEL_ID="gpt-4o-mini"

聊天完成客户端

OpenAIChatClient 使用聊天完成 API —— 这是最简单的选择,支持广泛的模型。

import asyncio
from agent_framework.openai import OpenAIChatClient

async def main():
    agent = OpenAIChatClient().as_agent(
        name="HelpfulAssistant",
        instructions="You are a helpful assistant.",
    )
    result = await agent.run("Hello, how can you help me?")
    print(result)

asyncio.run(main())

支持的工具: 函数工具、Web 搜索、本地 MCP 工具。

使用聊天辅助进行网页搜索

async def web_search_example():
    client = OpenAIChatClient()
    web_search = client.get_web_search_tool()

    agent = client.as_agent(
        name="SearchBot",
        instructions="You can search the web for current information.",
        tools=web_search,
    )
    result = await agent.run("What are the latest developments in AI?")
    print(result)

响应客户端

OpenAIResponsesClient 使用响应 API , 这是托管工具功能最丰富的选项。

import asyncio
from agent_framework.openai import OpenAIResponsesClient

async def main():
    agent = OpenAIResponsesClient().as_agent(
        name="FullFeaturedAgent",
        instructions="You are a helpful assistant with access to many tools.",
    )
    result = await agent.run("Write and run a Python script that calculates fibonacci numbers.")
    print(result)

asyncio.run(main())

支持的工具: 函数工具、工具审批、代码解释器、文件搜索、Web 搜索、托管 MCP、本地 MCP 工具。

包含响应客户端的托管工具

响应客户端为每个托管工具类型提供 get_*_tool() 方法:

async def hosted_tools_example():
    client = OpenAIResponsesClient()

    # Each tool is created via a client method
    code_interpreter = client.get_code_interpreter_tool()
    web_search = client.get_web_search_tool()
    file_search = client.get_file_search_tool(vector_store_ids=["vs_abc123"])
    mcp_tool = client.get_mcp_tool(
        name="GitHub",
        url="https://api.githubcopilot.com/mcp/",
        approval_mode="never_require",
    )

    agent = client.as_agent(
        name="PowerAgent",
        instructions="You have access to code execution, web search, files, and GitHub.",
        tools=[code_interpreter, web_search, file_search, mcp_tool],
    )
    result = await agent.run("Search the web for Python best practices, then write a summary.")
    print(result)

助手客户端

OpenAIAssistantProvider 使用助手 API - 具有内置代码解释器和文件搜索的服务器托管代理。 提供程序会自动管理助理生命周期。

import asyncio
from agent_framework.openai import OpenAIAssistantProvider
from openai import AsyncOpenAI

async def main():
    client = AsyncOpenAI()
    provider = OpenAIAssistantProvider(client)

    agent = await provider.create_agent(
        name="DataAnalyst",
        model="gpt-4o-mini",
        instructions="You analyze data using code execution.",
    )

    try:
        result = await agent.run("Calculate the first 20 prime numbers.")
        print(result)
    finally:
        await provider.delete_agent(agent.id)

asyncio.run(main())

支持的工具: 函数工具、代码解释器、文件搜索、本地 MCP 工具。

常见功能

这三种客户端类型都支持以下标准代理功能:

函数工具

from agent_framework import tool

@tool
def get_weather(location: str) -> str:
    """Get the weather for a given location."""
    return f"The weather in {location} is sunny, 25°C."

async def example():
    agent = OpenAIResponsesClient().as_agent(
        instructions="You are a weather assistant.",
        tools=get_weather,
    )
    result = await agent.run("What's the weather in Tokyo?")
    print(result)

多轮次对话

async def thread_example():
    agent = OpenAIResponsesClient().as_agent(
        instructions="You are a helpful assistant.",
    )
    session = await agent.create_session()

    result1 = await agent.run("My name is Alice", session=session)
    print(result1)
    result2 = await agent.run("What's my name?", session=session)
    print(result2)  # Remembers "Alice"

流媒体

async def streaming_example():
    agent = OpenAIResponsesClient().as_agent(
        instructions="You are a creative storyteller.",
    )
    print("Agent: ", end="", flush=True)
    async for chunk in agent.run("Tell me a short story about AI.", stream=True):
        if chunk.text:
            print(chunk.text, end="", flush=True)
    print()

使用代理

所有客户端类型都生成支持相同操作的标准 Agent

有关详细信息,请参阅 入门教程

后续步骤