C# .NET调用OpenAI及微软语音服务实现语音输入及输出

目前C#调用Open AI有很多的开源库，本文使用的nuget包如下：

 <PackageReference Include="DevExpressMvvm" Version="22.1.3" />
    <PackageReference Include="Microsoft.CognitiveServices.Speech" Version="1.29.0" />
    <PackageReference Include="OpenAI-DotNet" Version="6.8.7" />
    <PackageReference Include="WPF-UI" Version="2.0.3" />

都有开源库，所以实现起来就比较简单了，主要难度可能还是注册账号，有能力的可以自己试试，在这儿贴出部分调用的主要代码，项目已开源，Github搜索JamesBaiJun/OpenAI-Azure，记得点个Star。

调用ChatGPT使用Nuget的OpenAI-DotNet库，代码如下：

  static string key = string.Empty;
        static OpenAIClient api = new OpenAIClient(这里填你的Key);

        public static event Action<string> OnReceiveReply;
        public static event Action<string> OnReceivePartial;
        public static async Task AskAsync(List<Message> messages)
        {
            var chatRequest = new ChatRequest(messages, Model.GPT3_5_Turbo);
            await foreach (var result in api.ChatEndpoint.StreamCompletionEnumerableAsync(chatRequest))
            {
                foreach (var choice in result.Choices.Where(choice => choice.Delta?.Content != null))
                {
                    // Partial response content
                    //Debug.WriteLine(choice.Delta.Content);
                    OnReceivePartial?.Invoke(choice.Delta.Content);
                }

                foreach (var choice in result.Choices.Where(choice => choice.Message?.Content != null))
                {
                    // Completed response content
                    //Debug.WriteLine($"{choice.Message.Role}: {choice.Message.Content}");
                    OnReceiveReply?.Invoke(choice.Message.Content);
                }
            }
        }

软件截图如下，需要你填入自己的Key：