anbeime/skill / skills/digital-avatar-shopping-video/digital-avatar-shopping-video
digital-avatar-shopping-video
Verified电商运营与自媒体创作者在需要制作商品推荐或促销带货短视频时,使用此技能可一键生成“小省导购员”数字人口播视频。系统自动协同五大智能体完成脚本、画面与音画合成,产出适配抖音、快手等平台的成品,高效搞定带货视频制作。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 19.1 KB |
| assets/examples/example-output.md | markdown | 10.5 KB |
| references/agent-roles.md | markdown | 24.0 KB |
| references/content-templates.md | markdown | 20.9 KB |
| references/tech-integration.md | markdown | 20.4 KB |
| references/workflow-steps.md | markdown | 16.9 KB |
references/tech-integration.md
View on GitHub# 技术集成指南
## 目录
- [1. 技术方案概览](#1-技术方案概览)
- [2. 数字人API集成](#2-数字人api集成)
- [3. 视频合成工具集成](#3-视频合成工具集成)
- [4. 语音合成API集成](#4-语音合成api集成)
- [5. 端到端实现方案](#5-端到端实现方案)
- [6. 成本与性能考虑](#6-成本与性能考虑)
---
## 1. 技术方案概览
### MVP方案(最小可用版本)
**实现方式**:
- 使用智能体的图像生成能力生成关键帧
- 输出详细的视频合成方案和脚本
- 人工或使用基础视频编辑工具完成视频制作
**优点**:
- 无需额外API调用,成本最低
- 快速验证内容和流程
- 灵活性高,易于调整
**适用场景**:
- 内容策划和脚本验证
- 概念演示和原型制作
- 预算有限的项目
### 完整方案(技术集成版本)
**实现方式**:
- 集成数字人生成API(HeyGen、D-ID等)
- 集成视频合成工具(ffmpeg、剪映API等)
- 集成语音合成API(Azure TTS、百度TTS等)
- 实现端到端的自动化视频生成
**优点**:
- 完全自动化,批量生成效率高
- 视频质量稳定可控
- 可扩展性强
**适用场景**:
- 商业化运营
- 大批量视频生成
- 需要高效率的生产环境
---
## 2. 数字人API集成
### 2.1 HeyGen API
**API概述**:
HeyGen提供高质量的AI数字人视频生成服务,支持文本转视频、自定义数字人形象等功能。
**核心功能**:
- 文本转视频:输入文本,生成数字人讲解视频
- 自定义数字人:上传照片或视频,训练专属数字人
- 多语言支持:支持中英文等多种语言
- 实时生成:快速生成视频内容
**集成步骤**:
#### 步骤1:注册和获取API Key
1. 访问HeyGen官网注册账号
2. 进入开发者中心创建应用
3. 获取API Key
4. 在Skill中使用`skill_credentials`配置凭证
#### 步骤2:调用API生成数字人视频
```python
import os
from coze_workload_identity import requests
def generate_avatar_video(text: str, avatar_id: str, voice_id: str):
"""
使用HeyGen API生成数字人视频
参数:
text: 数字人台词
avatar_id: 数字人形象ID
voice_id: 语音ID
返回:
视频URL和下载链接
"""
# 获取凭证
skill_id = "7598200361526132770"
api_key = os.getenv("COZE_HEYGEN_API_KEY_" + skill_id)
if not api_key:
raise ValueError("缺少HeyGen API凭证")
# API端点
url = "https://api.heygen.com/v1/video.new"
headers = {
"Content-Type": "application/json",
"X-Api-Key": api_key
}
# 请求体
payload = {
"test": False, # 测试模式
"caption": False, # 是否显示字幕
"video_inputs": [
{
"character": {
"type": "avatar",
"avatar_id": avatar_id,
"avatar_style": "normal"
},
"voice": {
"type": "text",
"input_text": text,
"voice_id": voice_id
}
}
]
}
# 发起请求
try:
response = requests.post(url, headers=headers, json=payload, timeout=30)
response.raise_for_status()
data = response.json()
# 提取video_id
video_id = data.get("data", {}).get("video_id")
if not video_id:
raise Exception("未获取到video_id")
# 查询视频状态
status_url = f"https://api.heygen.com/v1/video.status?video_id={video_id}"
max_retries = 10
for i in range(max_retries):
status_response = requests.get(status_url, headers=headers, timeout=30)
status_response.raise_for_status()
status_data = status_response.json()
status = status_data.get("data", {}).get("status")
if status == "completed":
video_url = status_data.get("data", {}).get("video_url")
return {
"video_id": video_id,
"video_url": video_url,
"status": "completed"
}
elif status == "failed":
raise Exception("视频生成失败")
# 等待后重试
import time
time.sleep(5)
raise Exception("视频生成超时")
except requests.exceptions.RequestException as e:
raise Exception(f"HeyGen API调用失败: {str(e)}")
# 使用示例
if __name__ == "__main__":
# 生成单个场景的数字人视频
result = generate_avatar_video(
text="大家好,我是小省导购员!今天给大家推荐3款性价比超高的千元手机!",
avatar_id="your_avatar_id", # 从HeyGen获取
voice_id="your_voice_id" # 从HeyGen获取
)
print(f"视频生成成功: {result['video_url']}")
```
**脚本位置**:`scripts/heygen_avatar_generator.py`
#### 步骤3:配置凭证
调用`skill_credentials`工具配置HeyGen API密钥:
```python
skill_credentials(
credential_name="heygen_api",
auth_type=1,
allowed_domain="api.heygen.com",
env_variable_list=[
{
"variable_name": "API_KEY",
"api_key_location": 2, # Header
"api_key_param_name": "X-Api-Key",
"api_key_prefix": ""
}
]
)
```
### 2.2 D-ID API
**API概述**:
D-ID提供实时数字人生成服务,支持高质量的视频和动画。
**集成步骤**:
#### 步骤1:注册和获取API Key
1. 访问D-ID官网注册账号
2. 获取API Key
3. 在Skill中配置凭证
#### 步骤2:调用API生成数字人
```python
import os
from coze_workload_identity import requests
def generate_did_avatar(text: str, image_url: str):
"""
使用D-ID API生成数字人视频
参数:
text: 数字人台词
image_url: 数字人形象图片URL
返回:
视频URL
"""
skill_id = "7598200361526132770"
api_key = os.getenv("COZE_DID_API_KEY_" + skill_id)
if not api_key:
raise ValueError("缺少D-ID API凭证")
url = "https://api.d-id.com/talks"
headers = {
"Content-Type": "application/json",
"Authorization": f"Basic {api_key}"
}
payload = {
"script": {
"type": "text",
"input": text
},
"source_url": image_url,
"config": {
"fluent": True,
"pad_audio": 0.0
}
}
try:
response = requests.post(url, headers=headers, json=payload, timeout=30)
response.raise_for_status()
data = response.json()
talk_id = data.get("id")
if not talk_id:
raise Exception("未获取到talk_id")
# 查询视频状态
status_url = f"https://api.d-id.com/talks/{talk_id}"
max_retries = 10
for i in range(max_retries):
status_response = requests.get(status_url, headers=headers, timeout=30)
status_response.raise_for_status()
status_data = status_response.json()
status = status_data.get("status")
if status == "done":
result_url = status_data.get("result_url")
return {
"talk_id": talk_id,
"video_url": result_url,
"status": "done"
}
import time
time.sleep(5)
raise Exception("视频生成超时")
except requests.exceptions.RequestException as e:
raise Exception(f"D-ID API调用失败: {str(e)}")
```
**脚本位置**:`scripts/did_avatar_generator.py`
---
## 3. 视频合成工具集成
### 3.1 FFmpeg视频合成
**工具概述**:
FFmpeg是一个强大的音视频处理工具,可以用于视频剪辑、合并、添加水印等操作。
**集成方式**:
#### 使用Python调用FFmpeg
```python
import subprocess
import os
def composite_video(
avatar_video: str,
background_image: str,
product_images: list,
ui_elements: list,
output_path: str
):
"""
使用FFmpeg合成视频
参数:
avatar_video: 数字人视频文件路径
background_image: 背景图片文件路径
product_images: 产品图片列表
ui_elements: UI元素图片列表
output_path: 输出视频路径
"""
# FFmpeg命令模板
cmd = [
"ffmpeg",
"-i", avatar_video, # 数字人视频
"-i", background_image, # 背景图
# 添加产品图片和UI元素...
"-filter_complex", "[1:v]scale=1920:1080[bg];[0:v]scale=960:1080[avatar];[bg][avatar]overlay=(W-w)/2:0",
"-c:v", "libx264",
"-preset", "medium",
"-crf", "23",
"-pix_fmt", "yuv420p",
output_path
]
try:
subprocess.run(cmd, check=True, capture_output=True)
print(f"视频合成成功: {output_path}")
except subprocess.CalledProcessError as e:
raise Exception(f"FFmpeg执行失败: {e.stderr.decode()}")
# 使用示例
if __name__ == "__main__":
composite_video(
avatar_video="./avatar_scene1.mp4",
background_image="./background_scene1.png",
product_images=["./product1.png"],
ui_elements=["./ui_label1.png"],
output_path="./output_scene1.mp4"
)
```
**脚本位置**:`scripts/video_compositor_ffmpeg.py`
**依赖配置**(添加到SKILL.md):
```yaml
dependency:
system:
- apt-get update && apt-get install -y ffmpeg
```
### 3.2 剪映API集成
**API概述**:
剪映提供专业的视频编辑API,支持更高级的视频合成功能。
**集成步骤**:
```python
import os
from coze_workload_identity import requests
def create_jianying_project(materials: list):
"""
使用剪映API创建视频项目
参数:
materials: 素材列表(视频、图片、音频等)
"""
skill_id = "7598200361526132770"
api_key = os.getenv("COZE_JIANYING_API_KEY_" + skill_id)
if not api_key:
raise ValueError("缺少剪映API凭证")
url = "https://api.jianying.com/v1/projects"
headers = {
"Content-Type": "application/json",
"Authorization": f"Bearer {api_key}"
}
payload = {
"name": "数字人导购视频",
"materials": materials,
"timeline": {
"tracks": [
{
"type": "video",
"clips": [
{
"material_id": "avatar_video_id",
"start": 0,
"duration": 10
}
]
},
{
"type": "image",
"clips": [
{
"material_id": "background_id",
"start": 0,
"duration": 120
}
]
}
]
}
}
try:
response = requests.post(url, headers=headers, json=payload, timeout=30)
response.raise_for_status()
data = response.json()
project_id = data.get("project_id")
# 导出视频
export_url = f"https://api.jianying.com/v1/projects/{project_id}/export"
export_response = requests.post(export_url, headers=headers, timeout=30)
export_response.raise_for_status()
return export_response.json()
except requests.exceptions.RequestException as e:
raise Exception(f"剪映API调用失败: {str(e)}")
```
**脚本位置**:`scripts/jianying_compositor.py`
---
## 4. 语音合成API集成
### 4.1 Azure TTS
**API概述**:
Azure Text-to-Speech提供高质量的自然语音合成服务。
**集成步骤**:
```python
import os
from coze_workload_identity import requests
def text_to_speech(text: str, voice: str, output_path: str):
"""
使用Azure TTS生成语音
参数:
text: 待合成文本
voice: 语音类型(如zh-CN-XiaoxiaoNeural)
output_path: 输出音频文件路径
"""
skill_id = "7598200361526132770"
api_key = os.getenv("COZE_AZURE_TTS_KEY_" + skill_id)
if not api_key:
raise ValueError("缺少Azure TTS凭证")
url = f"https://eastus.tts.speech.microsoft.com/cognitiveservices/v1"
headers = {
"Ocp-Apim-Subscription-Key": api_key,
"Content-Type": "application/ssml+xml",
"X-Microsoft-OutputFormat": "audio-16khz-128kbitrate-mono-mp3"
}
ssml = f"""
<speak version='1.0' xml:lang='zh-CN'>
<voice xml:lang='zh-CN' name='{voice}'>
{text}
</voice>
</speak>
"""
try:
response = requests.post(url, headers=headers, data=ssml.encode('utf-8'), timeout=30)
response.raise_for_status()
with open(output_path, 'wb') as f:
f.write(response.content)
print(f"语音合成成功: {output_path}")
except requests.exceptions.RequestException as e:
raise Exception(f"Azure TTS调用失败: {str(e)}")
# 使用示例
if __name__ == "__main__":
text_to_speech(
text="大家好,我是小省导购员!今天给大家推荐3款性价比超高的千元手机!",
voice="zh-CN-XiaoxiaoNeural",
output_path="./audio_scene1.mp3"
)
```
**脚本位置**:`scripts/azure_tts.py`
### 4.2 百度TTS
```python
import os
from coze_workload_identity import requests
def baidu_text_to_speech(text: str, output_path: str):
"""
使用百度TTS生成语音
"""
skill_id = "7598200361526132770"
api_key = os.getenv("COZE_BAIDU_TTS_KEY_" + skill_id)
secret_key = os.getenv("COZE_BAIDU_TTS_SECRET_" + skill_id)
# 获取access_token
token_url = f"https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials&client_id={api_key}&client_secret={secret_key}"
token_response = requests.get(token_url, timeout=30)
token_data = token_response.json()
access_token = token_data.get("access_token")
# 调用TTS
tts_url = f"https://tsn.baidu.com/text2audio?tok={access_token}&tex={text}&cuid=user&ctp=1&lan=zh"
response = requests.get(tts_url, timeout=30)
with open(output_path, 'wb') as f:
f.write(response.content)
print(f"语音合成成功: {output_path}")
```
**脚本位置**:`scripts/baidu_tts.py`
---
## 5. 端到端实现方案
### 5.1 完整流程实现
```python
import os
from typing import List, Dict
class AvatarVideoGenerator:
"""数字人导购视频生成器(完整版)"""
def __init__(self, skill_id: str):
self.skill_id = skill_id
def generate_video(
self,
script_data: Dict,
visual_data: Dict,
avatar_data: Dict,
output_dir: str
) -> str:
"""
端到端生成数字人导购视频
参数:
script_data: 脚本数据
visual_data: 视觉设计数据
avatar_data: 数字人数据
output_dir: 输出目录
返回:
视频文件路径
"""
# 步骤1:生成数字人视频片段
avatar_videos = []
for scene in script_data["script_scenes"]:
avatar_video = self._generate_avatar_scene(
text=scene["dialogue"],
emotion=scene["avatar_expression"]
)
avatar_videos.append(avatar_video)
# 步骤2:合成背景和UI元素
background_video = self._composite_background(visual_data)
# 步骤3:合并所有元素
final_video = self._composite_all_elements(
avatar_videos=avatar_videos,
background_video=background_video,
output_path=os.path.join(output_dir, "final_video.mp4")
)
return final_video
def _generate_avatar_scene(self, text: str, emotion: str) -> str:
"""生成单个场景的数字人视频"""
# 调用数字人API
pass
def _composite_background(self, visual_data: Dict) -> str:
"""合成背景视频"""
# 使用FFmpeg或其他工具
pass
def _composite_all_elements(
self,
avatar_videos: List[str],
background_video: str,
output_path: str
) -> str:
"""合成所有元素"""
# 使用FFmpeg或其他工具
pass
# 使用示例
if __name__ == "__main__":
generator = AvatarVideoGenerator(skill_id="7598200361526132770")
# 读取脚本和设计数据
script_data = {...}
visual_data = {...}
avatar_data = {...}
# 生成视频
video_path = generator.generate_video(
script_data=script_data,
visual_data=visual_data,
avatar_data=avatar_data,
output_dir="./output"
)
print(f"视频生成完成: {video_path}")
```
**脚本位置**:`scripts/avatar_video_generator.py`
### 5.2 批量生成
```python
from avatar_video_generator import AvatarVideoGenerator
def batch_generate_videos(
product_lists: List[Dict],
output_dir: str
):
"""
批量生成导购视频
参数:
product_lists: 产品列表
output_dir: 输出目录
"""
generator = AvatarVideoGenerator(skill_id="7598200361526132770")
for i, products in enumerate(product_lists):
print(f"正在生成第{i+1}个视频...")
# 步骤1:内容策划
# 步骤2:脚本创作
# 步骤3:视觉设计
# 步骤4:数字人驱动
# 步骤5:视频合成
video_path = generator.generate_video(
script_data=script_data,
visual_data=visual_data,
avatar_data=avatar_data,
output_dir=os.path.join(output_dir, f"video_{i+1}")
)
print(f"视频{i+1}生成完成: {video_path}")
# 使用示例
if __name__ == "__main__":
product_lists = [
{"category": "手机", "products": [...]},
{"category": "服装", "products": [...]},
{"category": "家电", "products": [...]}
]
batch_generate_videos(product_lists, "./batch_output")
```
**脚本位置**:`scripts/batch_generator.py`
---
## 6. 成本与性能考虑
### 6.1 API成本分析
| 服务 | 免费额度 | 付费价格 | 适用场景 |
|-----|---------|---------|---------|
| HeyGen | 1个免费视频 | $29/月起 | 高质量数字人 |
| D-ID | 5个免费视频 | $29/月起 | 实时数字人 |
| Azure TTS | 500万字符/月 | $15/月起 | 高质量语音 |
| 百度TTS | 200万字符/天 | 按量付费 | 国内语音服务 |
**成本优化建议**:
- 优先使用免费额度
- 批量生成时选择合适的API套餐
- 考虑使用本地TTS降低成本
### 6.2 性能优化
**视频生成时间**:
- 数字人视频生成:每个场景30-60秒
- 视频合成:10-20秒/分钟视频
- 总计:100-120秒视频约需5-8分钟
**优化策略**:
- 并行生成多个数字人视频
- 使用缓存避免重复生成
- 预加载常用数字人形象
### 6.3 质量控制
**质量检查点**:
- 数字人表情是否自然
- 音画是否同步
- 视频质量是否达标
- 文件大小是否合理
**失败处理**:
- 重试机制:最多3次
- 降级方案:使用备用API
- 错误日志:记录失败原因
---
## 附录:推荐技术栈
### 入门级方案
- 数字人:D-ID(成本较低)
- 语音合成:百度TTS(国内,稳定)
- 视频合成:FFmpeg(开源免费)
- **总成本**:约$29/月
### 专业级方案
- 数字人:HeyGen(质量高)
- 语音合成:Azure TTS(自然度高)
- 视频合成:剪映API(功能强大)
- **总成本**:约$100/月
### 企业级方案
- 数字人:定制训练
- 语音合成:定制声音模型
- 视频合成:自建渲染服务
- **总成本**:根据需求定制
---
## 注意事项
1. **API凭证管理**:妥善保管API密钥,不要硬编码在脚本中
2. **配额限制**:注意API调用频率限制,避免超限
3. **版权问题**:确保使用的图片、音乐等素材有版权
4. **合规要求**:遵守各平台的API使用规范
5. **备份策略**:重要数据及时备份
6. **错误监控**:建立完善的错误监控和日志系统
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |