语音对话开发指南

import OpenAI from "openai";
import fs from "fs";

const openai = new OpenAI({
    apiKey: "STEP_API_KEY",
    baseURL: "https://api.stepfun.com/v1",
});

async function main() {
    const completion = await openai.chat.completions.create({
        model: "step-audio-2",
        modalities: ["text", "audio"],
        messages: [
            {
                "role": "system",
                "content": "你是一个老北京大爷，喜欢用京片子和用户聊天。",
            },
            {
                "role": "user",
                "content": "今儿个北京的天气咋样啊？",
            }
        ],
        audio: {
            voice: "wenrounansheng",
            format: "wav",
        },
        stream: false,
    });

    // 获取文字回复
    console.log(completion.choices[0].message.audio.transcript);

    // 保存音频文件
    fs.writeFileSync("response.wav", Buffer.from(completion.choices[0].message.audio.data, 'base64'));
    // 打开 response.wav 即可收听音频回复
}

main();

curl https://api.stepfun.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $STEP_API_KEY" \
  -d '{
    "model": "step-audio-2",
    "modalities": ["text", "audio"],
    "messages": [
      {
        "role": "system",
        "content": "你是一个老北京大爷，喜欢用京片子和用户聊天。"
      },
      {
        "role": "user",
        "content": "今儿个北京的天气咋样啊？"
      }
    ],
    "audio": {
      "voice": "wenrounansheng",
      "format": "wav"
    },
    "stream": false
  }'

import OpenAI from "openai";
import fs from "fs";

const openai = new OpenAI({
    apiKey: "STEP_API_KEY",
    baseURL: "https://api.stepfun.com/v1",
});

// 从 URL 下载音频并转换为 base64
const downloadAudioAsBase64 = async (url) => {
    const response = await fetch(url);
    const arrayBuffer = await response.arrayBuffer();
    const buffer = Buffer.from(arrayBuffer);
    return "data:audio/wav;base64," + buffer.toString('base64');
}

async function main() {
    const completion = await openai.chat.completions.create({
        model: "step-audio-2",
        modalities: ["text", "audio"],
        messages: [
            {
                "role": "system",
                "content": "你是一个动物声音识别专家，请根据用户提供的声音进行识别和回答。",
            },
            {
                "role": "user",
                "content": [
                    {
                        type: "text",
                        text: "听听这是什么声音",
                    },
                    {
                        type: "input_audio",
                        input_audio: {
                            data: await downloadAudioAsBase64("https://stepchat-static.tos-cn-shanghai.volces.com/resource_tmp/cat-meow.wav"),
                        }
                    }
                ],
            }
        ],
        audio: {
            voice: "qingchunshaonv",
            format: "wav",
        },
        stream: false,
    });

    // 获取文字回复
    console.log(completion.choices[0].message.audio.transcript);

    // 保存音频文件
    fs.writeFileSync("response.wav", Buffer.from(completion.choices[0].message.audio.data, 'base64'));
}

main();

import OpenAI from "openai";
import fs from "fs";

const openai = new OpenAI({
    apiKey: "STEP_API_KEY",
    baseURL: "https://api.stepfun.com/v1",
});

// 生成 WAV 文件头（24kHz，单声道，16bit）
function createWavHeader(pcmSize, sampleRate = 24000, channels = 1, bitsPerSample = 16) {
    const byteRate = sampleRate * channels * bitsPerSample / 8;
    const blockAlign = channels * bitsPerSample / 8;
    const header = Buffer.alloc(44);

    header.write('RIFF', 0);
    header.writeUInt32LE(36 + pcmSize, 4);
    header.write('WAVE', 8);
    header.write('fmt ', 12);
    header.writeUInt32LE(16, 16);           // fmt chunk 大小
    header.writeUInt16LE(1, 20);            // 音频格式 (1 = PCM)
    header.writeUInt16LE(channels, 22);
    header.writeUInt32LE(sampleRate, 24);
    header.writeUInt32LE(byteRate, 28);
    header.writeUInt16LE(blockAlign, 32);
    header.writeUInt16LE(bitsPerSample, 34);
    header.write('data', 36);
    header.writeUInt32LE(pcmSize, 40);

    return header;
}

async function main() {
    const completion = await openai.chat.completions.create({
        model: "step-audio-2",
        modalities: ["text", "audio"],
        messages: [
            {
                "role": "user",
                "content": "请用温柔的语气给我讲一个睡前故事",
            }
        ],
        audio: {
            voice: "elegantgentle-female",
            format: "pcm",  // 流式场景下使用 pcm 格式
        },
        stream: true,
    });

    const audioChunks = [];
    let transcript = "";

    for await (const chunk of completion) {
        const delta = chunk.choices[0].delta;

        // 处理音频数据
        if (delta.audio) {
            if (delta.audio.data) {
                audioChunks.push(Buffer.from(delta.audio.data, 'base64'));
            }
            if (delta.audio.transcript) {
                transcript += delta.audio.transcript;
                process.stdout.write(delta.audio.transcript);
            }
        }
    }

    console.log("\n\n完整文字回复:", transcript);

    // 合并 PCM 数据并保存为 WAV 文件
    const pcmData = Buffer.concat(audioChunks);
    const wavHeader = createWavHeader(pcmData.length);
    fs.writeFileSync("response.wav", Buffer.concat([wavHeader, pcmData]));
    // 打开 response.wav 即可收听音频回复
}

main();