ai-video-mcp
Turns text scripts into complete videos with stock footage, AI narration, and subtitles via MCP tools for AI assistants.
README
🎬 AI Video MCP Server
Turn text scripts into complete videos with stock footage, AI narration, and subtitles. Standard MCP tool — works with Claude, OpenClaw, Cursor, and any MCP-compatible AI assistant.
✨ Features
- 📝 Script-driven — Just provide text, auto-split into chapters
- 🎥 Smart footage — Auto-matches stock videos from Pexels per chapter
- 🗣️ TTS narration — Minimax voice synthesis, 6 voice options
- 📺 Auto subtitles — Generates SRT subtitles from narration
- 🎬 One-click compose — ffmpeg auto-merges into final video
- 🔌 MCP standard — Compatible with all MCP-enabled AI assistants
🚀 Quick Start
1. Install
pip install ai-video-mcp
Or from source:
git clone https://github.com/yqyqstart-crypto/ai-video-mcp.git
cd ai-video-mcp
pip install -e .
2. Set API Keys
export MINIMAX_API_KEY="your-minimax-key"
export PEXELS_API_KEY="your-pexels-key"
| Key | Get it from | Purpose |
|---|---|---|
| MINIMAX_API_KEY | platform.minimaxi.com | TTS voice synthesis |
| PEXELS_API_KEY | pexels.com/api | Stock video search |
3. Use
Command Line
ai-video-mcp
MCP Config (Claude Desktop / OpenClaw / Cursor)
{
"mcpServers": {
"ai-video": {
"command": "ai-video-mcp",
"env": {
"MINIMAX_API_KEY": "your-key",
"PEXELS_API_KEY": "your-key"
}
}
}
}
Once configured, just tell the AI:
"Generate a video from this script: /path/to/script.txt"
The AI will call the generate_video tool automatically.
📖 Tools
generate_video
Generate a complete video from a text script.
Parameters:
| Param | Type | Default | Description |
|---|---|---|---|
script_path |
string | required | Path to script file |
output_dir |
string | out |
Output directory |
voice |
string | male-qn-qingse |
TTS voice ID |
speed |
float | 1.05 |
Speech speed (0.5-2.0) |
resolution |
string | 1920x1080 |
Output resolution |
video_duration |
int | 10 |
Pexels clip duration per chapter (sec) |
Returns:
{
"success": true,
"final_video": "out/final_video.mp4",
"chapters": 3,
"details": [
{
"chapter": 0,
"title": "Chapter 1",
"query": "search keywords",
"video": "out/chapter_00/chapter.mp4",
"tts": true
}
]
}
list_voices
List available TTS voice IDs.
list_styles
List available video styles (planned).
📝 Script Format
Plain text (.txt) — Split with ---
Artificial intelligence is changing our lives. From smartphones to self-driving cars, AI is everywhere.
---
In healthcare, AI helps doctors diagnose diseases more accurately.
---
In the future, AI will continue to drive social progress.
Markdown (.md) — Split with ##
## The Rise of AI
Artificial intelligence is changing our lives...
## Healthcare Revolution
In healthcare, AI helps doctors...
## Future Outlook
In the future, AI will continue...
🎙️ Available Voices
| ID | Name | Gender |
|---|---|---|
male-qn-qingse |
Youth | Male |
female-shaonv |
Girl | Female |
male-qn-jingying |
Elite Male | Male |
female-yujie |
Elegant Lady | Female |
presenter male |
Male Host | Male |
presenter female |
Female Host | Female |
📂 Output Structure
out/
├── final_video.mp4 # ← Final complete video
├── concat.txt # ffmpeg merge list
├── chapter_00/
│ ├── source.mp4 # Pexels raw footage
│ ├── audio.mp3 # TTS narration audio
│ ├── audio.srt # Subtitle file
│ └── chapter.mp4 # Composed chapter video
├── chapter_01/
│ └── ...
└── chapter_02/
└── ...
🏗️ Workflow
┌─────────────┐
│ Text Script │
│ .txt / .md │
└──────┬──────┘
│
┌──────▼──────┐
│Parse Chapters│
└──────┬──────┘
│
┌────────────┼────────────┐
│ │ │
┌──────▼──────┐ ┌──▼────┐ ┌─────▼─────┐
│ Pexels Search│ │ TTS │ │ Subtitles │
│ Stock Videos │ │ Voice │ │ SRT │
└──────┬──────┘ └──┬────┘ └─────┬─────┘
│ │ │
└────────────┼────────────┘
│
┌──────▼──────┐
│ffmpeg Compose│
│Video+Audio+Sub│
└──────┬──────┘
│
┌──────▼──────┐
│ final.mp4 │
└─────────────┘
🔧 Requirements
- Python 3.8+
- ffmpeg (system install)
- requests
- fastmcp
💡 Why MCP?
MCP (Model Context Protocol) lets AI assistants call external tools directly. With this MCP Server installed:
- ✅ Claude Desktop can generate videos
- ✅ OpenClaw can generate videos
- ✅ Cursor can generate videos
- ✅ Any MCP-enabled AI assistant works
No coding needed. No command line. Just tell the AI.
📄 License
MIT
🤝 Contributing
PRs welcome! Especially:
- More TTS engines (Edge TTS, Azure TTS)
- More video sources (Pixabay, Coverr)
- AI image generation (DALL-E, Stable Diffusion)
- Custom subtitle styles
- Background music mixing
⭐ Star this repo
If you find this useful, give it a ⭐ to help others discover it!
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。