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.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.