MCP Notes Server
A beginner-friendly MCP server for managing personal notes. Enables Claude to create, list, read, search, update, and delete notes saved as Markdown files.
README
MCP Notes Server — Your First MCP Server with Claude (Python)
A beginner-friendly, fully reproducible tutorial for building a Model Context Protocol (MCP) server in Python and connecting it to Claude.
The example server is a personal notes manager: Claude can create, list, read, search, update, and delete notes that are saved as Markdown files on your own computer.
📖 New to this? Follow the complete step-by-step guide in
TUTORIAL.md. It explains every single command and shows the exact output you should see.
What is MCP?
MCP (Model Context Protocol) is an open standard that lets AI apps like Claude talk to external programs. You write a small server exposing:
- Tools — actions Claude can take (e.g. "save a note")
- Resources — read-only data Claude can pull in (e.g. "all my notes")
Claude is the client: when you ask it something, it can decide to call your tools. Build the server once, and any MCP-aware app can use it.
Do I need to write a client too? Usually no — the client is an existing app like Claude Desktop or Claude Code. You only write a server. This repo includes an optional
client.pypurely to show what a client does under the hood; seeTUTORIAL.md§9.
Quickstart
Requires uv (the tutorial shows how to install it). Then, from this folder:
# 1. Set up the environment (installs Python 3.12 + the MCP SDK, pinned)
uv sync
# 2. Test the server in the visual MCP Inspector
uv run mcp dev notes_server.py
To connect it to Claude Desktop, add this to your
claude_desktop_config.json (use the absolute path from pwd):
{
"mcpServers": {
"notes": {
"command": "uv",
"args": ["--directory", "/ABSOLUTE/PATH/TO/mcp_tutorial", "run", "notes_server.py"]
}
}
}
To connect it to Claude Code:
claude mcp add notes -- uv --directory "$(pwd)" run notes_server.py
Full details, expected output, and troubleshooting are in
TUTORIAL.md.
Project structure
| File | Purpose |
|---|---|
notes_server.py |
The MCP server: 6 tools + 1 resource |
client.py |
Optional standalone MCP client — chat with the server from your terminal (no Claude Desktop needed) |
pyproject.toml |
Project metadata, requires Python ≥ 3.10, depends on mcp[cli] and anthropic |
uv.lock |
Exact pinned versions of every dependency (commit this!) |
.python-version |
Pins the Python interpreter to 3.12 for reproducibility |
.env.example |
Template for your API key — copy to .env (git-ignored) and fill in |
TUTORIAL.md |
The complete step-by-step walkthrough |
.gitignore |
Excludes the virtual env and your personal notes from git |
The tools
| Tool | What it does |
|---|---|
add_note(title, body) |
Create a new note |
update_note(title, body) |
Replace an existing note's content |
list_notes() |
List all note titles |
read_note(title) |
Read one note |
search_notes(query) |
Find notes by title or content |
delete_note(title) |
Delete a note permanently |
Plus a resource, notes://all, that returns every note concatenated.
Tested with
- uv 0.11.19
- Python 3.12.13 (pinned via
.python-version) - mcp 1.27.2
Because uv.lock and .python-version are committed, anyone who runs
uv sync gets the exact same environment.
License
MIT — see LICENSE. Use it, fork it, teach with 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。