todo-mcp
Enables AI assistants to manage Microsoft To Do tasks via the Graph API, supporting multiple accounts, encrypted auth, and full CRUD operations.
README
todo-mcp
An MCP (Model Context Protocol) server that gives AI assistants access to Microsoft To Do via the Microsoft Graph API. Supports multiple accounts (personal and work), encrypted token storage, and all common task operations.
Features
- List task lists and tasks across any number of accounts
- Create, update, complete, and delete tasks
- Get all pending tasks across every list in one call
- Multi-account support (e.g. separate
workandpersonalaliases) - Refresh tokens stored encrypted with AES-256-GCM on disk
- Non-blocking device code authentication flow (URL returned immediately)
Prerequisites
- Node.js 18 or later
- An Azure app registration with the Microsoft Graph
Tasks.Read,Tasks.ReadWrite, andoffline_accessdelegated permissions (see below)
1. Create an Azure App Registration
- Go to portal.azure.com → Microsoft Entra ID → App registrations → New registration
- Name it anything (e.g.
My Todo MCP) - Supported account types: choose "Accounts in any organizational directory and personal Microsoft accounts" (enables both work and personal accounts)
- No redirect URI needed — click Register
- Copy the Application (client) ID — you'll need it below
- Go to API permissions → Add a permission → Microsoft Graph → Delegated
- Add:
Tasks.Read,Tasks.ReadWrite,offline_access - Click Grant admin consent (or users will be prompted on first sign-in)
- Add:
- Go to Authentication → Advanced settings → enable "Allow public client flows" (required for device code)
- Go to Manifest → set
"requestedAccessTokenVersion": 2in theapisection → Save
2. Install
git clone <repo-url> todo-mcp
cd todo-mcp
npm install
3. Configure
Option A — Claude Desktop / Cowork (claude_desktop_config.json)
Add this to %APPDATA%\Claude\claude_desktop_config.json (Windows) or ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):
{
"mcpServers": {
"todo-mcp": {
"command": "node",
"args": ["C:/path/to/todo-mcp/index.js"],
"env": {
"TODO_MCP_CLIENT_ID": "<your-client-id>",
"TODO_MCP_MASTER_KEY": "<random-base64-key>"
}
}
}
}
Generate a secure master key (PowerShell):
[System.Convert]::ToBase64String([System.Security.Cryptography.RandomNumberGenerator]::GetBytes(32))
Or on macOS/Linux:
openssl rand -base64 32
Option B — VS Code (.vscode/mcp.json)
The included .vscode/mcp.json prompts for the master key securely on first use. Set TODO_MCP_CLIENT_ID in your environment or add it to the env block.
4. Authenticate Accounts
Once the server is running, use the authenticate_account tool from your AI assistant. It returns a device code URL immediately:
authenticate_account(account="work", tenant_id="<your-org-tenant-id>")
authenticate_account(account="personal", tenant_id="consumers")
Go to the URL shown, enter the code, and sign in. The token is saved automatically in the background.
For personal Microsoft accounts (Outlook, Hotmail, Live), always pass tenant_id="consumers".
For work/school accounts, pass your organization's tenant ID (a UUID), or "organizations" to accept any.
Environment Variables
| Variable | Required | Description |
|---|---|---|
TODO_MCP_MASTER_KEY |
Yes | Base64 key used to encrypt stored refresh tokens |
TODO_MCP_CLIENT_ID |
Yes* | Azure app client ID (*or hardcode in auth.js) |
TODO_MCP_TENANT_ID |
No | Default tenant ID (defaults to "common") |
TODO_MCP_DEFAULT_ACCOUNT |
No | Default account alias (defaults to "default") |
TODO_MCP_CONFIG_PATH |
No | Override path for the config file (defaults to ~/.todo-mcp-config.json) |
Available Tools
| Tool | Description |
|---|---|
list_accounts |
List all stored account aliases |
authenticate_account |
Start device code auth for a new or existing account |
set_default_account |
Change which account is used when none is specified |
delete_account |
Remove a stored account |
list_todo_lists |
List all task lists for an account |
get_tasks |
Get tasks from a specific list (filterable by status) |
get_all_pending_tasks |
Get all non-completed tasks across every list |
create_task |
Create a task with optional due date, importance, and body |
update_task |
Update a task's title, due date, importance, or body |
complete_task |
Mark a task as completed |
delete_task |
Delete a task |
All data tools accept an optional account parameter to target a specific alias. If omitted, the default account is used.
Token Storage
Refresh tokens are stored in ~/.todo-mcp-config.json encrypted with AES-256-GCM using a key derived from TODO_MCP_MASTER_KEY. The file is created with mode 0600 (owner read/write only). Never commit this file.
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。