Agents-flex为 Java 构建可靠的 AI Agent

OpenAIImageModelConfig config = new OpenAIImageModelConfig();
config.setApiKey(System.getenv("OPENAI_API_KEY"));
ImageModel model = new OpenAIImageModel(config);

GenerateImageRequest request = new GenerateImageRequest();
request.setPrompt("雨后的未来城市，电影感光影");
request.setSize(1024, 1024);

ImageResponse response = model.generate(request);
response.getImages().get(0)
    .writeToFile(new File("output/city.png"));

AliyunTextToSpeechConfig config = new AliyunTextToSpeechConfig();
config.setAppKey(System.getenv("ALIYUN_APP_KEY"));
config.setAccessKeyId(System.getenv("ALIYUN_ACCESS_KEY_ID"));
config.setAccessKeySecret(System.getenv("ALIYUN_ACCESS_KEY_SECRET"));

TextToSpeechModel model = new AliyunTextToSpeechModel(config);
TextToSpeechRequest request = new TextToSpeechRequest(
    "欢迎使用 Agents-Flex 多模态能力"
);
TextToSpeechResponse response = model.tts(request);
response.writeTo(new File("output/reply.mp3"));

AliyunSpeechToTextConfig config = new AliyunSpeechToTextConfig();
config.setAppKey(System.getenv("ALIYUN_APP_KEY"));
config.setAccessKeyId(System.getenv("ALIYUN_ACCESS_KEY_ID"));
config.setAccessKeySecret(System.getenv("ALIYUN_ACCESS_KEY_SECRET"));

SpeechToTextModel model = new AliyunSpeechToTextModel(config);
SpeechToTextRequest request = new SpeechToTextRequest();
request.setAudioFile(new File("meeting.mp3"));

SpeechToTextResponse response = model.stt(request);
System.out.println(response.getResult());

AliyunWanVideoModelConfig config = new AliyunWanVideoModelConfig();
config.setApiKey(System.getenv("DASHSCOPE_API_KEY"));
AliyunWanVideoModel model = new AliyunWanVideoModel(config);

GenerateVideoRequest request = new GenerateVideoRequest();
request.setPrompt("纸飞机飞过日出时的未来城市");
request.setSize(1280, 720);
request.setDuration(5);

VideoResponse response = model.generateAndWait(request);
response.getVideo().writeToFile(
    new File("output/city.mp4")
);