Todoist MCP Server
Enables AI assistants to manage Todoist tasks, projects, and labels through natural language. It provides a comprehensive suite of tools for task organization, productivity tracking, and structured workflows like daily planning.
README
Todoist MCP Server
An MCP (Model Context Protocol) server that provides comprehensive task management through Todoist. This enables AI assistants like Claude to create, manage, and organize tasks using natural language.
Features
Tools (21 total)
| Tool | Description |
|---|---|
list_tasks |
List tasks with filtering (by project, priority, due date, etc.) |
get_task |
Get detailed task info including comments and subtasks |
create_task |
Create tasks with natural language due dates |
update_task |
Update task content, due date, priority, labels |
complete_task |
Mark tasks complete (handles recurring) |
uncomplete_task |
Reopen completed tasks |
delete_task |
Permanently delete tasks |
move_task |
Move tasks between projects/sections |
list_projects |
List all projects |
create_project |
Create new projects |
update_project |
Update project name, color, view style |
delete_project |
Permanently delete projects |
list_sections |
List sections within projects |
create_section |
Create sections |
update_section |
Rename or collapse/expand sections |
move_section |
Move sections between projects |
delete_section |
Permanently delete sections |
list_labels |
List all labels |
create_label |
Create new labels |
update_label |
Update label name, color, order |
delete_label |
Permanently delete labels |
add_comment |
Add comments to tasks |
get_productivity_stats |
Get completion statistics |
quick_add |
Natural language task entry |
Resources
| Resource | Description |
|---|---|
todoist://today |
Today's tasks including overdue |
todoist://upcoming |
Tasks due in the next 7 days |
todoist://overdue |
All overdue tasks |
todoist://projects |
Project overview with task counts |
todoist://labels |
Labels and usage counts |
Prompts
| Prompt | Description |
|---|---|
daily_planning |
Morning planning workflow |
weekly_review |
GTD-style weekly review |
quick_capture |
Rapid multi-task capture |
task_breakdown |
Break large tasks into subtasks |
Setup
1. Get Your Todoist API Token
- Go to https://todoist.com/app/settings/integrations/developer
- Scroll to "API token"
- Copy your token
2. Install Dependencies
cd todoist-mcp
npm install
3. Build
npm run build
4. Add to Claude Code
claude mcp add --transport stdio todoist \
--env TODOIST_API_TOKEN=your_token_here \
-- node /path/to/todoist-mcp/dist/index.js
Or add manually to your Claude Code configuration:
{
"mcpServers": {
"todoist": {
"command": "node",
"args": ["/path/to/todoist-mcp/dist/index.js"],
"env": {
"TODOIST_API_TOKEN": "your_token_here"
}
}
}
}
5. Verify Connection
In Claude Code, type /mcp to check the server is connected.
Usage Examples
Create a Task
"Add a task to review the budget report by Friday"
The AI will use create_task with natural language due date parsing.
View Today's Tasks
"What do I need to do today?"
Uses the todoist://today resource or list_tasks with dueToday: true.
Daily Planning
"Let's do my daily planning"
Invokes the daily_planning prompt for a guided planning session.
Quick Capture
"Add these tasks: call dentist tomorrow, review PR by EOD, update docs"
Uses quick_add or create_task for natural language entry.
Initialize GTD Structure (One-Time Setup)
Run the setup script to create a GTD-inspired structure:
TODOIST_API_TOKEN=your_token npx tsx scripts/setup-gtd-structure.ts
This creates:
- Projects: Work, Personal (with sub-projects), Recurring, Someday/Maybe, Reference
- Labels: Context (@computer, @phone), Energy (@high-energy), Time (@5min, @deep-work), Status (@waiting-for)
Natural Language Due Dates
Todoist supports flexible date parsing:
| Input | Result |
|---|---|
today |
Today |
tomorrow |
Tomorrow |
tomorrow at 2pm |
Tomorrow at 2:00 PM |
next monday |
Next Monday |
in 3 days |
3 days from now |
jan 15 |
January 15 |
every day |
Daily recurring |
every monday at 9am |
Weekly on Monday at 9 AM |
every weekday |
Monday through Friday |
Priority Mapping
| API Value | Display | Color |
|---|---|---|
4 |
P1 (highest) | Red |
3 |
P2 | Orange |
2 |
P3 | Yellow |
1 |
P4 (default) | None |
Rate Limiting
The server implements rate limiting to stay within Todoist's API limits:
- Limit: 450 requests per 15 minutes
- Caching: Projects and labels are cached for 5 minutes
Development
Run in Development Mode
npm run dev
This runs TypeScript in watch mode.
Test with MCP Inspector
npm run inspect
Opens the MCP Inspector for testing tools and resources.
Project Structure
todoist-mcp/
├── src/
│ ├── index.ts # Entry point
│ ├── types/ # TypeScript type definitions
│ ├── services/ # Todoist client, caching, rate limiting
│ ├── tools/ # MCP tool implementations
│ ├── resources/ # MCP resource implementations
│ └── prompts/ # MCP prompt implementations
├── dist/ # Compiled output
├── package.json
└── tsconfig.json
Troubleshooting
"TODOIST_API_TOKEN environment variable is required"
Make sure you've set the token in your MCP configuration:
claude mcp add --env TODOIST_API_TOKEN=your_token ...
Rate Limit Errors
The server automatically handles rate limiting, but if you hit limits:
- Wait 15 minutes for the window to reset
- Check
get_productivity_statsto see current rate limit usage
Tasks Not Appearing
- Verify your API token is correct
- Check that tasks aren't filtered out (completed, different project, etc.)
- Use
list_taskswithout filters to see all tasks
Quick Add Project Assignment
The quick_add tool uses Todoist's natural language parser, which may not always parse project names with spaces correctly. For example, "Task #My Project" might not assign to the "My Project" project.
Workaround: Use create_task with explicit projectId for reliable project assignment, or use project names without spaces.
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。