bell-mcp
A local MCP server that rings a bell through your speakers to provide audible notifications when AI completes tasks or needs attention.
README
bell-mcp
A local MCP server that rings a bell through your speakers — so you get an audible nudge when your AI assistant finishes a task or needs your attention.
Four built-in sounds, synthesized from scratch with the Python standard library (no audio assets to download):
| id | name | sound |
|---|---|---|
beep |
Retro Beep | harsh 90s PC-speaker square-wave beep |
ding |
Ding | single soft struck-bell tone |
ding-ding |
Double Ding | two struck-bell tones in a row |
warning |
Warning Alarm | urgent alternating two-tone alarm |
You can preset a default sound and/or pick a sound per call.
Requirements
- Python ≥ 3.11 and
uv - An audio player on PATH:
- macOS:
afplay(built in) - Linux:
paplay,aplay, orffplay - Windows: PowerShell (built in)
- macOS:
Tools
ring_bell(sound?, message?)— ring a bell.soundis one ofbeep/ding/ding-ding/warning; if omitted, the preset default is used.messageis an optional note (e.g."build finished") echoed back in the confirmation.list_bells()— list available sounds (id, name, description) and show the current default.
Configuration
The default bell (used when ring_bell is called without sound) is set via the
BELL_DEFAULT_SOUND environment variable. Valid values: beep, ding, ding-ding,
warning. If unset or invalid, it falls back to ding.
Install
Clone the repository, then register the server with your MCP client.
git clone git@github.com:taylor224/bell-mcp.git
cd bell-mcp
uv sync
Claude Code
Run this from the cloned project directory:
claude mcp add bell -e BELL_DEFAULT_SOUND=ding -- uv run --directory "$(pwd)" bell-mcp
Add --scope user to make it available in every project (instead of just the current one).
Verify it connected:
claude mcp list
Other MCP clients (Claude Desktop, etc.)
Add an entry to your client's MCP config. Replace /absolute/path/to/bell-mcp
with the path where you cloned this repo:
{
"mcpServers": {
"bell": {
"command": "uv",
"args": ["run", "--directory", "/absolute/path/to/bell-mcp", "bell-mcp"],
"env": { "BELL_DEFAULT_SOUND": "ding" }
}
}
}
Once registered, ask your assistant to ring the bell — e.g. "ring ding-ding whenever you finish a task" or "ring the warning alarm if you need my input."
Development
uv sync # install dependencies
uv run bell-mcp # run the server over stdio
uv run python -m bell_mcp.sounds # (re)generate the WAV files
The WAV files live in bell_mcp/assets/ and are generated automatically on first run.
To tweak the timbres, edit the frequency / decay / partial values in
bell_mcp/sounds.py and regenerate.
Project layout
bell_mcp/
server.py # FastMCP server + tools, cross-platform playback
sounds.py # WAV synthesis (stdlib only); auto-generated on first run
assets/ # generated *.wav files
pyproject.toml # bell-mcp entry point
License
MIT
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。