Drama Generator
Tagline: 短剧生成 - 从剧本到视频,一键生成 | From Script to Video, One-Click Generation
中文 / Chinese
简介
Drama Generator(短剧生成器) - 自动化短剧视频生成系统
将剧本文本自动转换为多角色配音的专业短剧视频。支持场景切换、角色对话、旁白等完整短剧元素。
核心功能
- 多角色对话 - 自动识别角色,分配不同声音(6种声音选择)
- 剧本解析 - 支持标准短剧格式(场景 + 角色 + 对话)
- 智能配音 - 基于角色自动选择合适的 TTS 声音
- 场景切换 - 美观的场景转场动画效果
- 对话框样式 - 专业的对话呈现,角色名突出显示
- 旁白支持 - 特殊的旁白视觉风格(引号样式)
- 完全自动化 - 一键从剧本到视频的完整流水线
使用场景
- 短剧创作 - 快速生成短剧视频内容
- 内容创作 - 制作对话类视频素材
- 教育培训 - 场景化教学演示视频
- 营销推广 - 故事化产品介绍视频
- 娱乐创作 - AI 辅助短剧制作
剧本格式
支持标准的短剧剧本格式:
[场景N - 地点 - 时间]
角色名: 对话内容
旁白: 旁白内容
示例:
[场景1 - 办公室 - 早晨]
旁白: 又是忙碌的一天。
小明: 早啊,今天的会议准备好了吗?
小红: 准备好了!
[场景2 - 会议室 - 上午]
老板: 大家好,我们开始吧。
小明: (紧张) 好的,老板。
快速开始
安装
# npm 全局安装
npm install -g openclaw-drama-generator
# 或从 GitHub 克隆
git clone https://github.com/ZhenRobotics/openclaw-drama-generator.git
cd openclaw-drama-generator
npm install
配置 API Key
# OpenAI(推荐)
export OPENAI_API_KEY="sk-..."
# 或其他提供商
export AZURE_TTS_KEY="..."
export ALIYUN_ACCESS_KEY_ID="..."
生成第一个短剧
# 1. 创建剧本文件
cat > my-drama.txt <<'EOF'
[场景1 - 咖啡厅 - 下午]
旁白: 这是一个温暖的午后。
小美: 你好,好久不见!
小王: 是啊,最近很忙。
小美: 在学习做短剧视频,用AI生成的。
小王: 听起来很酷!
EOF
# 2. 一键生成视频
drama-generator my-drama.txt
# 3. 查看结果
mpv out/my-drama.mp4
技术特点
- 多提供商支持 - OpenAI、Azure、阿里云、腾讯云 TTS
- 自动声音映射 - 智能分配角色声音(6种声音)
- 精确时间戳 - Whisper API 同步对话
- 专业视觉 - Remotion 渲染引擎,赛博风格
- 高性能 - 本地渲染,快速输出
输出规格
- 分辨率: 1080 x 1920 (竖屏,适合抖音/快手/视频号)
- 帧率: 30 fps
- 格式: MP4 (H.264 + AAC)
- 音频: 多角色配音,自动拼接
- 时长: 根据剧本自动计算
角色声音
| 声音 | 特点 | 适合角色 |
|---|---|---|
| alloy | 中性,清晰 | 男性角色 |
| echo | 男性,稳重 | 成熟男性 |
| fable | 英式,优雅 | 旁白 |
| onyx | 深沉,有力 | 领导/老板 |
| nova | 女性,活泼 | 年轻女性 |
| shimmer | 柔和,温暖 | 旁白/女性 |
成本估算
使用 OpenAI TTS + Whisper:
- 每分钟短剧约 $0.015
- 非常经济实惠
性能
处理时间(1分钟短剧):
- 剧本解析: <1秒
- TTS 生成: 10-20秒
- 时间戳提取: 5-10秒
- 场景生成: <1秒
- 视频渲染: 30-90秒
- 总计: 50-120秒
项目链接
- GitHub: https://github.com/ZhenRobotics/openclaw-drama-generator
- npm: https://www.npmjs.com/package/openclaw-drama-generator
- 文档: 完整的中英文文档
- 示例: 查看 scripts/examples/office-drama.txt
English
Introduction
Drama Generator - Automated drama video generation system
Automatically converts drama scripts into multi-character voiced professional drama videos. Supports scene transitions, character dialogues, narration, and complete drama elements.
Core Features
- Multi-Character Dialogue - Automatic character recognition with different voices (6 voice options)
- Script Parsing - Support for standard drama format (scenes + characters + dialogues)
- Smart Voice Assignment - Automatically select appropriate TTS voices based on characters
- Scene Transitions - Beautiful scene transition animations
- Dialogue Box Style - Professional dialogue presentation with highlighted character names
- Narration Support - Special visual style for narration (quotation marks style)
- Full Automation - One-click complete pipeline from script to video
Use Cases
- Drama Creation - Quickly generate drama video content
- Content Creation - Create dialogue-based video materials
- Education & Training - Scenario-based teaching demonstration videos
- Marketing - Story-based product introduction videos
- Entertainment - AI-assisted drama production
Script Format
Standard drama script format:
[Scene N - Location - Time]
Character Name: Dialogue content
Narrator: Narration content
Example:
[Scene 1 - Office - Morning]
Narrator: Another busy day begins.
Ming: Good morning, is the meeting ready?
Hong: Ready!
[Scene 2 - Meeting Room - Morning]
Boss: Hello everyone, let's begin.
Ming: (nervous) Yes, boss.
Quick Start
Installation
# npm global install
npm install -g openclaw-drama-generator
# Or clone from GitHub
git clone https://github.com/ZhenRobotics/openclaw-drama-generator.git
cd openclaw-drama-generator
npm install
Configure API Key
# OpenAI (recommended)
export OPENAI_API_KEY="sk-..."
# Or other providers
export AZURE_TTS_KEY="..."
export ALIYUN_ACCESS_KEY_ID="..."
Generate Your First Drama
# 1. Create script file
cat > my-drama.txt <<'EOF'
[Scene 1 - Cafe - Afternoon]
Narrator: A warm afternoon.
Mei: Hello, long time no see!
Wang: Yes, been very busy.
Mei: Learning to make drama videos with AI.
Wang: Sounds cool!
EOF
# 2. Generate video with one command
drama-generator my-drama.txt
# 3. View result
mpv out/my-drama.mp4
Technical Features
- Multi-Provider Support - OpenAI, Azure, Alibaba Cloud, Tencent Cloud TTS
- Auto Voice Mapping - Smart character voice assignment (6 voices)
- Precise Timestamps - Whisper API dialogue synchronization
- Professional Visuals - Remotion rendering engine, cyber style
- High Performance - Local rendering, fast output
Output Specifications
- Resolution: 1080 x 1920 (Portrait, suitable for TikTok/Douyin/Kuaishou)
- Frame Rate: 30 fps
- Format: MP4 (H.264 + AAC)
- Audio: Multi-character voices, automatically concatenated
- Duration: Auto-calculated from script
Character Voices
| Voice | Characteristics | Suitable For |
|---|---|---|
| alloy | Neutral, clear | Male characters |
| echo | Male, steady | Mature males |
| fable | British, elegant | Narration |
| onyx | Deep, powerful | Leaders/bosses |
| nova | Female, lively | Young females |
| shimmer | Soft, warm | Narration/females |
Cost Estimation
Using OpenAI TTS + Whisper:
- About $0.015 per minute of drama
- Very cost-effective
Performance
Processing time (1-minute drama):
- Script parsing: <1s
- TTS generation: 10-20s
- Timestamp extraction: 5-10s
- Scene generation: <1s
- Video rendering: 30-90s
- Total: 50-120s
Project Links
- GitHub: https://github.com/ZhenRobotics/openclaw-drama-generator
- npm: https://www.npmjs.com/package/openclaw-drama-generator
- Documentation: Complete bilingual documentation
- Examples: See scripts/examples/office-drama.txt
License: MIT-0 Version: 2.0.0 Author: ZhenStaff