HeyGemWeb

KO | EN

gitlite — search

#javascript #python #hacktoberfest #ai #typescript #llm #rust #deep-learning #windows #awesome #cli #vue #awesome-list #open-source

(★ 19)

一个基于HeyGem接口开发的ai数字人web应用

HeyGemWeb Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# HeyGem · AI 数字人创作平台 / HeyGem · AI Digital Human Creation Platform

[中文](#heygem--ai-数字人创作平台) | [English](#heygem--ai-digital-human-creation-platform-1)

## HeyGem · AI 数字人创作平台

HeyGem是一个基于AI技术的数字人视频创作平台，用户可以通过输入文本快速生成具有专业质量的数字人视频。

### 项目特色

- 🎤 AI语音克隆：上传音频样本，快速克隆属于你的独特声音
- 🧑 数字人形象：上传视频创建个性化的数字人形象
- 🎬 一键生成：输入文本即可自动生成数字人视频
- 🖼️ 现代化界面：基于React和TailwindCSS构建的美观用户界面
- 🚀 高性能：采用FastAPI后端，提供流畅的用户体验

### 项目结构

```
HeyGemWeb/
├── web/
│   └── app.py          # 主程序文件
├── img/                # 截图和图片资源目录
├── backend/
│   └── docker-compose.yaml  # 后端服务编排文件
└── README.md           # 项目说明文档
```

### 技术架构

#### 整体架构
HeyGem采用前后端一体化的设计，使用FastAPI作为后端框架，前端使用React构建单页应用(SPA)并直接嵌入到HTML响应中。项目通过调用多个AI服务容器实现完整的数字人视频生成流程。

#### 核心组件
1. **主应用** ([app.py](file:///D:/GUOSHIYIN/github%E4%BB%A3%E7%A0%81/HeyGemWeb/web/app.py#L1-L59)):
   - 基于FastAPI构建的Web服务器
   - 内嵌React前端应用
   - 提供RESTful API接口
   - 处理数据持久化

2. **AI服务容器**:
   - TTS服务 (端口 18180)：基于Fish-Speech的文本转语音服务
   - F2F服务 (端口 8383)：Face2Face数字人面部动画生成服务
   - ASR服务 (端口 10095)：自动语音识别服务

#### 工作流程
1. 用户在前端界面输入文本内容
2. 文本通过API发送到后端服务
3. 后端调用TTS服务将文本转换为语音
4. 语音数据传递给F2F服务生成面部动画
5. 最终生成的视频返回给前端展示

### 核心功能

#### 1. 创作工作室 (Creation Studio)
- 文本脚本编辑
- 声音模型选择
- 数字人形象选择
- 音频合成与视频生成
- 实时预览功能

#### 2. 素材库 (Materials Library)
- 声音模型管理 (上传、训练、删除)
- 数字人形象管理 (上传、预览、删除)
- 支持音频文件上传与处理
- 支持视频文件上传与缩略图生成

#### 3. 我的作品 (My Works)
- 视频作品展示与管理
- 视频预览功能
- 作品删除功能

### 使用技术

#### 后端技术栈
- [FastAPI](https://fastapi.tiangolo.com/): 高性能Python Web框架，用于构建API
- [Uvicorn](https://www.uvicorn.org/): ASGI服务器，用于运行FastAPI应用
- [Requests](https://docs.python-requests.org/): Python HTTP库，用于与AI服务通信
- [MoviePy](https://zulko.github.io/moviepy/): 视频编辑库，用于处理视频文件和生成缩略图
- [Pydantic](https://docs.pydantic.dev/): 数据验证和设置管理库

#### 前端技术栈
- [React](https://reactjs.org/): JavaScript库，用于构建用户界面
- [Tailwind CSS](https://tailwindcss.com/): CSS框架，用于快速UI开发
- [Font Awesome](https://fontawesome.com/): 图标库
- [Google Fonts](https://fonts.google.com/): 字体服务

#### 数据存储
- JSON文件存储：用于持久化用户数据（声音模型、数字人形象、生成的视频等）

### 界面展示

以下是项目的主要界面截图：

<div align="center">
  <img src="img/1.png" alt="创作工作室界面" width="45%" />
  <img src="img/2.png" alt="素材库界面" width="45%" />
  <p>创作工作室和素材库界面</p>
</div>

<div align="center">
  <img src="img/3.png" alt="数字人形象管理" width="45%" />
  <img src="img/4.png" alt="我的作品界面" width="45%" />
  <p>数字人形象管理和我的作品界面</p>
</div>

<div align="center">
  <img src="img/5.png" alt="声音模型管理" width="45%" />
  <img src="img/6.png" alt="视频预览界面" width="45%" />
  <p>声音模型管理和视频预览界面</p>
</div>

### 运行环境要求

- Docker & Docker Compose
- Python 3.8+
- NVIDIA GPU (用于AI模型推理)
- 至少16GB内存

### 快速开始

1. 启动AI服务容器：
```bash
cd backend
docker-compose up -d
```

2. 安装Python依赖并启动主应用：
```bash
pip install -r requirements.txt
cd web
python app.py
```

3. 在浏览器中访问 `http://localhost:8000`

### 使用说明

1. **创作工作室**：输入文本脚本，选择声音模型和数字人形象，生成视频
2. **素材库**：上传和管理声音模型与数字人形象
3. **我的作品**：查看和管理已生成的视频作品

### 依赖服务

本项目依赖以下Docker容器服务：
- TTS服务 (端口 18180)：基于Fish-Speech的文本转语音服务
- F2F服务 (端口 8383)：Face2Face数字人面部动画生成服务
- ASR服务 (端口 10095)：自动语音识别服务

## HeyGem · AI Digital Human Creation Platform

HeyGem is an AI-powered digital human video creation platform that allows users to quickly generate professional-quality digital human videos by simply inputing text.

### Key Features

- 🎤 AI Voice Cloning: Upload audio samples to quickly clone your unique voice
- 🧑 Digital Human Avatars: Upload videos to create personalized digital human avatars
- 🎬 One-click Generation: Automatically generate digital human videos from text input
- 🖼️ Modern Interface: Beautiful user interface built with React and TailwindCSS
- 🚀 High Performance: FastAPI backend for smooth user experience

### Project Structure

```
HeyGemWeb/
├── web/
│   └── app.py          # Main application file
├── img/                # Screenshot and image resources directory
├── backend/
│   └── docker-compose.yaml  # Backend service orchestration file
└── README.md           # Project documentation
```

### Technical Architecture

#### Overall Architecture
HeyGem adopts an integrated front-end and back-end design, using FastAPI as the back-end framework and React to build a single-page application (SPA) embedded directly into the HTML response. The project implements a complete digital human video generation process by calling multiple AI service containers.

#### Core Components
1. **Main Application** ([app.py](file:///D:/GUOSHIYIN/github%E4%BB%A3%E7%A0%81/HeyGemWeb/web/app.py#L1-L59)):
   - Web server built on FastAPI
   - Embedded React front-end application
   - Provides RESTful API interfaces
   - Handles data persistence

2. **AI Service Containers**:
   - TTS service (port 18180): Text-to-speech service based on Fish-Speech
   - F2F service (port 8383): Face2Face digital human facial animation generation service
   - ASR service (port 10095): Automatic speech recognition service

#### Workflow
1. Users input text content in the front-end interface
2. Text is sent to the back-end service via API
3. Back-end calls TTS service to convert text to speech
4. Audio data is passed to F2F service to generate facial animation
5. The final generated video is returned to the front-end for display

### Core Features

#### 1. Creation Studio
- Text script editing
- Voice model selection
- Digital human avatar selection
- Audio synthesis and video generation
- Real-time preview function

#### 2. Materials Library
- Voice model management (upload, training, delete)
- Digital human avatar management (upload, preview, delete)
- Audio file upload and processing support
- Video file upload and thumbnail generation support

#### 3. My Works
- Video works display and management
- Video preview function
- Work deletion function

### Technologies Used

#### Backend Stack
- [FastAPI](https://fastapi.tiangolo.com/): High-performance Python web framework for building APIs
- [Uvicorn](https://www.uvicorn.org/): ASGI server for running FastAPI applications
- [Requests](https://docs.python-requests.org/): Python HTTP library for communicating with AI services
- [MoviePy](https://zulko.github.io/moviepy/): Video editing library for processing video files and generating thumbnails
- [Pydantic](https://docs.pydantic.dev/): Data validation and settings management library

#### Frontend Stack
- [React](https://reactjs.org/): JavaScript library for building user interfaces
- [Tailwind CSS](https://tailwindcss.com/): CSS framework for rapid UI development
- [Font Awesome](https://fontawesome.com/): Icon library
- [Google Fonts](https://fonts.google.com/): Font service

#### Data Storage
- JSON file storage: Used for persisting user data (voice models, digital human avatars, generated videos, etc.)

### Interface Showcase

Here are the main interface screenshots of the project:

<div align="center">
  <img src="img/1.png" alt="Creation Studio Interface" width="45%" />
  <img src="img/2.png" alt="Materials Library Interface" width="45%" />
  <p>Creation Studio and Materials Library Interface</p>
</div>

<div align="center">
  <img src="img/3.png" alt="Digital Human Avatar Management" width="45%" />
  <img src="img/4.png" alt="My Works Interface" width="45%" />
  <p>Digital Human Avatar Management and My Works Interface</p>
</div>

<div align="center">
  <img src="img/5.png" alt="Voice Model Management" width="45%" />
  <img src="img/6.png" alt="Video Preview Interface" width="45%" />
  <p>Voice Model Management and Video Preview Interface</p>
</div>

### Requirements

- Docker & Docker Compose
- Python 3.8+
- NVIDIA GPU (for AI model inference)
- At least 16GB RAM

### Quick Start

1. Start AI service containers:
```bash
cd backend
docker-compose up -d
```

2. Install Python dependencies and start the main application:
```bash
pip install -r requirements.txt
cd web
python app.py
```

3. Access the application in your browser at `http://localhost:8000`

### Usage Guide

1. **Creation Studio**: Input text scripts, select voice models and digital human avatars to generate videos
2. **Materials Library**: Upload and manage voice models and digital human avatars
3. **My Works**: View and manage generated video works

### Dependency Services

This project depends on the following Docker container services:
- TTS service (port 18180): Text-to-speech service based on Fish-Speech
- F2F service (port 8383): Face2Face digital human facial animation generation service
- ASR service (port 10095): Automatic speech recognition service

---

💖 支持作者
如果你觉得这个项目对你有帮助，欢迎通过下方二维码赞赏支持作者的持续开发！

![赞赏码](https://raw.githubusercontent.com/zcc135820-web/HeyGemWeb/main/img/dashang.png)

---

⚠️ 免责声明
本软件仅供学习和研究使用。使用者需要：

- 遵守相关法律法规
- 尊重知识产权
- 不得用于商业用途
- 对使用本软件产生的任何后果自行承担责任

⭐ 如果这个项目对你有帮助，请给个Star支持一下！

💡 有商业化想法？欢迎交流合作，共同探索AI写作的无限可能！

HeyGemWeb

HeyGemWeb Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

HeyGemWeb

HeyGemWeb Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README