Muzaic MCP Server
Enables generating AI-powered soundtracks for videos via Muzaic AI from any MCP client.
README
🎵 Muzaic MCP Server
Official MCP server for Muzaic AI — generate custom music for video content
<!-- mcp-name: io.github.muzaic-ai/muzaic-mcp -->
✨ What is this?
Muzaic MCP Server connects the Muzaic AI music generation API to any LLM or agent that speaks the Model Context Protocol (MCP). Generate AI-powered soundtracks for videos — directly from Claude, Cursor, VS Code, or any MCP client.
- 🎶 Generate single tracks or multi-region soundtracks
- 🏷️ Browse 30+ music tags (style, mood, genre, cultural)
- 🎛️ Fine-tune with 5 parameters: intensity, tempo, rhythm, tone, variance
- 📈 Dynamic keyframes for parameter changes over time
- ✅ Tag compatibility validation
- 💰 Check account balance and token usage
⚠️ Note: Muzaic tokens are needed to generate music. Get your API key from adminpanel.muzaic.ai.
📦 Quick Install
pip install muzaic-mcp
Or with uv:
uv pip install muzaic-mcp
🏃 Quick Start
export MUZAIC_API_KEY=your_api_key_here
muzaic-mcp
🔧 Configuration
Claude Desktop
Go to Claude → Settings → Developer → Edit Config and add:
{
"mcpServers": {
"Muzaic": {
"command": "uvx",
"args": ["muzaic-mcp"],
"env": {
"MUZAIC_API_KEY": "<your-api-key>"
}
}
}
}
Cursor
Windows: Add to %USERPROFILE%\.cursor\mcp.json (e.g., C:\Users\YourName\.cursor\mcp.json)
macOS/Linux: Add to ~/.cursor/mcp.json
{
"mcpServers": {
"Muzaic": {
"command": "uvx",
"args": ["muzaic-mcp"],
"env": {
"MUZAIC_API_KEY": "<your-api-key>"
}
}
}
}
Note: If uvx is not in your PATH, use the full path. On Windows with Python installed, you can find it with:
where.exe uvx
# Or if using uv:
where.exe uv
# Then use: "C:\\path\\to\\uv.exe" with args: ["tool", "run", "muzaic-mcp"]
Alternatively, if you have Python installed, you can use:
{
"mcpServers": {
"Muzaic": {
"command": "python",
"args": ["-m", "muzaic_mcp.server"],
"env": {
"MUZAIC_API_KEY": "<your-api-key>"
}
}
}
}
Claude Code
claude mcp add-json "Muzaic" '{"command":"uvx","args":["muzaic-mcp"],"env":{"MUZAIC_API_KEY":"<your-api-key>"}}'
VS Code
Create .vscode/mcp.json in your workspace:
{
"servers": {
"Muzaic": {
"command": "uvx",
"args": ["muzaic-mcp"],
"env": {
"MUZAIC_API_KEY": "<your-api-key>"
}
}
}
}
🛠️ Available Tools
| Tool | Description | Cost |
|---|---|---|
muzaic_get_tags |
List all available music tags with descriptions | Free |
muzaic_validate_tags |
Check tag combination compatibility | Free |
muzaic_account_info |
Check token balance and usage | Free |
muzaic_generate_music |
Generate a single AI music track | 1 token/sec |
muzaic_create_soundtrack |
Create multi-region soundtrack with different styles | 1 token/sec |
muzaic_regenerate |
Regenerate music from a previous hash | 1 token/sec |
💬 Example Prompts
Try these with Claude, Cursor, or any MCP client:
"What music styles can you generate with Muzaic?""Create a 60-second upbeat pop track for a product video""Generate a cinematic soundtrack that starts calm and builds to epic over 2 minutes""Make a 30-second chill ambient track for a meditation app intro""Create a soundtrack with 3 sections: calm intro, energetic middle, peaceful outro"
🎛️ Music Parameters
All parameters range from 1 (low) to 9 (high):
| Parameter | 1 | 9 |
|---|---|---|
| intensity | Calm, ambient | High energy, powerful |
| tempo | Very slow | Very fast |
| rhythm | Simple, steady | Complex patterns |
| tone | Dark, moody | Bright, cheerful |
| variance | Repetitive | Diverse, changing |
Dynamic Keyframes
All parameters except tempo support keyframes — values that change over the duration of the track. Positions are expressed as percentages (0–100%) of the track duration.
# 30-second track: intensity drops at 10s, then rises to max at the end
intensity: [[0, 5], [33, 1], [100, 9]]
↑ start at 5 ↑ drop to 1 at 33% (=10s) ↑ rise to 9 at 100% (=30s)
Tempo only accepts a static value (1–9) — keyframes are not supported for tempo.
🐛 Troubleshooting
API key not set
Error: MUZAIC_API_KEY is not set
Make sure the MUZAIC_API_KEY environment variable is configured in your MCP client settings.
uvx not found
Windows:
# Find uvx location
where.exe uvx
# If not found, install uv first:
# pip install uv
# Or download from: https://github.com/astral-sh/uv/releases
# Then use full path in mcp.json:
# "command": "C:\\Users\\YourName\\AppData\\Local\\Programs\\uv\\uv.exe"
# "args": ["tool", "run", "muzaic-mcp"]
macOS/Linux:
# Find uvx location
which uvx
# If not found, install uv:
# curl -LsSf https://astral.sh/uv/install.sh | sh
# Then use full path in config:
# "command": "/usr/local/bin/uvx"
Alternative: Use Python directly if uvx is unavailable:
{
"mcpServers": {
"Muzaic": {
"command": "python",
"args": ["-m", "muzaic_mcp.server"],
"env": {
"MUZAIC_API_KEY": "<your-api-key>"
}
}
}
}
Timeout on long tracks
Music generation can take up to 5 minutes for 20-minute tracks. This is normal. If using MCP Inspector in dev mode, you may see timeout errors even though the generation completes.
🧪 Development
# Clone
git clone https://github.com/MUZAIC-Organization/muzaic-mcp.git
cd muzaic-mcp
# Install with dev dependencies
uv sync
# Or with pip:
# pip install -e ".[dev]"
# Run in dev mode
# Windows PowerShell:
$env:MUZAIC_API_KEY="your_key"
# macOS/Linux:
export MUZAIC_API_KEY=your_key
uv run mcp dev muzaic_mcp/server.py
# Or with pip:
# python -m muzaic_mcp.server
# Run tests
uv run pytest
# Or with pip:
# pytest
Local Development Setup
If you're developing locally and want to test without publishing to PyPI:
Windows:
# Install in editable mode
pip install -e .
# Test the command
$env:MUZAIC_API_KEY="your_key"
muzaic-mcp
macOS/Linux:
# Install in editable mode
pip install -e .
# Test the command
export MUZAIC_API_KEY=your_key
muzaic-mcp
Testing with MCP Inspector
npx @modelcontextprotocol/inspector
# → Transport: stdio
# → Command: uv run muzaic-mcp
# Or: python -m muzaic_mcp.server
Windows Note: If you encounter issues with npx, ensure Node.js is installed and in your PATH. You can also use the full path to npx if needed.
📄 License
MIT — see LICENSE for details.
🔗 Links
- Website: muzaic.ai
- API Docs: docs.muzaic.ai
- Get API Key: adminpanel.muzaic.ai
- Support: service@muzaic.ai
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。