babybuddy-mcp
Connect AI assistants to your Baby Buddy instance to log and query feedings, sleep, diaper changes, and more.
README
babybuddy-mcp
An MCP server for the Baby Buddy baby tracking application. Connect AI assistants to your babybuddy instance to log and query feedings, sleep, diaper changes, and more.
Built with FastMCP and designed to run alongside babybuddy in Docker.
Requirements
- mise (for local development)
- Python 3.13+
- A running babybuddy instance with API access
Setup
1. Get your API token
- Log in to your babybuddy web UI
- Click your username in the top-right corner → Settings
- Copy the API Key shown on the page
2. Configure environment
cp .env.example .env
Edit .env:
BABYBUDDY_URL=http://localhost:8000 # URL of your babybuddy instance
BABYBUDDY_TOKEN=your-api-key-here # Token from step 1
MCP_TRANSPORT=http # "http" for Docker, "stdio" for Claude Desktop
MCP_PORT=8080
Running with Docker
docker compose up --build
The MCP server will be available at http://localhost:8080/mcp/.
If babybuddy is running on the same Docker network, set BABYBUDDY_URL to use its service name (e.g. http://babybuddy:8000).
Connecting to Claude Desktop (stdio mode)
Set MCP_TRANSPORT=stdio in .env, then add to your Claude Desktop claude_desktop_config.json:
{
"mcpServers": {
"babybuddy": {
"command": "mise",
"args": ["x", "python@3.13", "--", "uv", "run", "python", "-m", "babybuddy_mcp"],
"cwd": "/path/to/babybuddy-mcp",
"env": {
"BABYBUDDY_URL": "http://localhost:8000",
"BABYBUDDY_TOKEN": "your-api-key-here",
"MCP_TRANSPORT": "stdio"
}
}
}
}
Connecting via HTTP transport
Add to your MCP client config:
{
"mcpServers": {
"babybuddy": {
"url": "http://localhost:8080/mcp/"
}
}
}
Available Tools
Tools are organized by domain with namespaced names (e.g. feedings_list_feedings).
| Domain | Tools |
|---|---|
| children | list_children, get_child, create_child, update_child |
| diapers | list_diaper_changes, create_diaper_change, update_diaper_change, delete_diaper_change |
| feedings | list_feedings, create_feeding, update_feeding, delete_feeding |
| sleep | list_sleep, create_sleep, update_sleep, delete_sleep |
| pumping | list_pumping, create_pumping, update_pumping, delete_pumping |
| tummy_times | list_tummy_times, create_tummy_time, update_tummy_time, delete_tummy_time |
| timers | list_timers, get_timer, create_timer, update_timer, delete_timer |
| measurements | list/create/update/delete × bmi, height, weight, head_circumference, temperature |
| notes | list_notes, create_note, update_note, delete_note, list_tags, create_tag, update_tag, delete_tag |
Environment Variables
| Variable | Required | Default | Description |
|---|---|---|---|
BABYBUDDY_URL |
Yes | — | URL of your babybuddy instance |
BABYBUDDY_TOKEN |
Yes | — | API token from babybuddy Settings |
MCP_TRANSPORT |
No | http |
Transport: http or stdio |
MCP_HOST |
No | 0.0.0.0 |
Host to bind (HTTP transport only) |
MCP_PORT |
No | 8080 |
Port to listen on (HTTP transport only) |
REQUEST_TIMEOUT |
No | 30.0 |
Timeout in seconds for babybuddy API calls |
DEFAULT_PAGE_SIZE |
No | 100 |
Max records returned per list call |
Development
# Install tools and dependencies
mise install
mise run install
# Run tests
mise run test
# Lint
mise run lint
# Type check
mise run typecheck
Example prompts
- "Log a feeding for Alice — she had breast milk from the left breast from 2pm to 2:25pm today"
- "How many times did Alice eat yesterday?"
- "Record a diaper change for Alice at 3pm — wet only"
- "Start a sleep timer for Alice"
- "Alice just woke up, end her sleep timer and log it"
- "What was Alice's weight at her last checkup?"
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。