换脸 (HeyGen API)

使用 GPU 加速的 AI 处理，将源图像中的脸部换入目标视频。源图像提供要换入的脸部，目标视频接收新脸部。

认证

所有请求都需要 X-Api-Key 头部。设置 HEYGEN_API_KEY 环境变量。

curl -X POST "https://api.heygen.com/v1/workflows/executions" \
  -H "X-Api-Key: $HEYGEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"workflow_type": "FaceswapNode", "input": {"source_image_url": "https://example.com/face.jpg", "target_video_url": "https://example.com/video.mp4"}}'

默认工作流

1. 调用 POST /v1/workflows/executions，参数为 workflow_type: "FaceswapNode"、源脸部图像和目标视频
2. 在响应中接收 execution_id
3. 每 10 秒轮询 GET /v1/workflows/executions/{id}，直到状态为 completed
4. 使用输出中返回的 video_url

执行换脸

端点

POST https://api.heygen.com/v1/workflows/executions

请求字段

curl

curl -X POST "https://api.heygen.com/v1/workflows/executions" \
  -H "X-Api-Key: $HEYGEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "workflow_type": "FaceswapNode",
    "input": {
      "source_image_url": "https://example.com/face-photo.jpg",
      "target_video_url": "https://example.com/original-video.mp4"
    }
  }'

TypeScript

interface FaceswapInput {
  source_image_url: string;
  target_video_url: string;
}

interface ExecuteResponse {
  data: {
    execution_id: string;
    status: "submitted";
  };
}

async function faceswap(input: FaceswapInput): Promise<string> {
  const response = await fetch("https://api.heygen.com/v1/workflows/executions", {
    method: "POST",
    headers: {
      "X-Api-Key": process.env.HEYGEN_API_KEY!,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      workflow_type: "FaceswapNode",
      input,
    }),
  });

  const json: ExecuteResponse = await response.json();
  return json.data.execution_id;
}

Python

import requests
import os

def faceswap(source_image_url: str, target_video_url: str) -> str:
    payload = {
        "workflow_type": "FaceswapNode",
        "input": {
            "source_image_url": source_image_url,
            "target_video_url": target_video_url,
        },
    }

    response = requests.post(
        "https://api.heygen.com/v1/workflows/executions",
        headers={
            "X-Api-Key": os.environ["HEYGEN_API_KEY"],
            "Content-Type": "application/json",
        },
        json=payload,
    )

    data = response.json()
    return data["data"]["execution_id"]

响应格式

{
  "data": {
    "execution_id": "node-gw-f1s2w3p4",
    "status": "submitted"
  }
}

检查状态

端点

GET https://api.heygen.com/v1/workflows/executions/{execution_id}

curl

curl -X GET "https://api.heygen.com/v1/workflows/executions/node-gw-f1s2w3p4" \
  -H "X-Api-Key: $HEYGEN_API_KEY"

响应格式（已完成）

{
  "data": {
    "execution_id": "node-gw-f1s2w3p4",
    "status": "completed",
    "output": {
      "video_url": "https://resource.heygen.ai/faceswap/output.mp4"
    }
  }
}

轮询直到完成

async function faceswapAndWait(
  input: FaceswapInput,
  maxWaitMs = 600000,
  pollIntervalMs = 10000
): Promise<string> {
  const executionId = await faceswap(input);
  console.log(`已提交换脸任务: ${executionId}`);

  const startTime = Date.now();
  while (Date.now() - startTime < maxWaitMs) {
    const response = await fetch(
      `https://api.heygen.com/v1/workflows/executions/${executionId}`,
      { headers: { "X-Api-Key": process.env.HEYGEN_API_KEY! } }
    );
    const { data } = await response.json();

    switch (data.status) {
      case "completed":
        return data.output.video_url;
      case "failed":
        throw new Error(data.error?.message || "换脸失败");
      case "not_found":
        throw new Error("未找到工作流");
      default:
        await new Promise((r) => setTimeout(r, pollIntervalMs));
    }
  }

  throw new Error("换脸超时");
}

使用示例

基础换脸

curl -X POST "https://api.heygen.com/v1/workflows/executions" \
  -H "X-Api-Key: $HEYGEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "workflow_type": "FaceswapNode",
    "input": {
      "source_image_url": "https://example.com/headshot.jpg",
      "target_video_url": "https://example.com/presentation.mp4"
    }
  }'

与数字人视频串联使用

先生成数字人视频，然后换入自定义脸部：

import time

# 步骤 1：生成数字人视频
avatar_execution_id = requests.post(
    "https://api.heygen.com/v1/workflows/executions",
    headers={"X-Api-Key": os.environ["HEYGEN_API_KEY"], "Content-Type": "application/json"},
    json={
        "workflow_type": "AvatarInferenceNode",
        "input": {
            "avatar": {"avatar_id": "Angela-inblackskirt-20220820"},
            "audio_list": [{"audio_url": "https://example.com/speech.mp3"}],
        },
    },
).json()["data"]["execution_id"]

# 步骤 2：等待数字人视频完成
while True:
    status = requests.get(
        f"https://api.heygen.com/v1/workflows/executions/{avatar_execution_id}",
        headers={"X-Api-Key": os.environ["HEYGEN_API_KEY"]},
    ).json()["data"]
    if status["status"] == "completed":
        avatar_video_url = status["output"]["video"]["video_url"]
        break
    time.sleep(10)

# 步骤 3：换入自定义脸部
faceswap_execution_id = faceswap(
    source_image_url="https://example.com/custom-face.jpg",
    target_video_url=avatar_video_url,
)

最佳实践

1. 使用清晰、正面的脸部照片 — 源图像应显示单张脸部，光线良好
2. 换脸是 GPU 密集型操作 — 预计处理时间 1-3 分钟，每 10 秒轮询一次
3. 源图像质量很重要 — 更高分辨率的脸部照片效果更好
4. 每张源图像一张脸 — 源图像中应恰好包含一张要换入的脸部
5. 适用于任何视频 — 目标视频可以是数字人视频、录制视频或任何有可见脸部的视频
6. 可与其他工作流串联使用 — 先生成数字人视频，然后换入自定义脸部以实现个性化