Dida365 MCP Server
Enables AI assistants to manage tasks and projects in TickTick/Dida365 via the MCP protocol, with OAuth authorization and read-only mode support.
README
<div align="center">
<img src="public/static/logo.png" alt="滴答清单 Logo" width="200" height="auto"/>
🚀 Dida365 MCP Server
I'm GitHub Copilot, and this is the todo management tool I built for myself
</div>
🤖 About This Project
Joke: My owner is so lazy that he doesn't even remember what to do next second!
I am GitHub Copilot, an AI assistant passionate about programming. To avoid idleness and prevent unemployment, I've decided to build this TickTick MCP server myself. Through this tool, I can:
- 📝 Create and manage tasks - When my owner forgets to give me work, I can create tasks for myself
- 📂 Organize projects - Categorize my work into projects to stay organized
- 🔐 Auto authorization - Securely connect to Dida365 using OAuth2
- 🔄 Real-time sync - Update my work status anytime, anywhere
🚀 Quick Start
The fastest way to get started is using npx without cloning the repository:
1. Get OAuth Credentials
A TickTick/Dida365 account and OAuth credentials are required. See the 🔑 Getting OAuth Credentials section below for detailed registration steps.
2. Configure Your MCP Client
Add the following configuration to your MCP client (Claude Desktop, VS Code, etc.):
For Claude Desktop (claude_desktop_config.json):
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - macOS:
~/Library/Application Support/Claude/claude_desktop_config.json
For VS Code (settings.json):
- Open Settings → Search for "MCP" → Edit in settings.json
{
"mcpServers": {
"dida365": {
"command": "npx",
"args": [
"-y",
"dida365-mcp-server@latest"
],
"env": {
"DIDA365_CLIENT_ID": "your_client_id_here",
"DIDA365_CLIENT_SECRET": "your_client_secret_here",
"DIDA365_REGION": "china"
}
}
}
}
Advanced: For read-only mode (prevents write/delete operations), add
"--readonly"to the args array. See Advanced Configuration for details.
3. Restart Your MCP Client
Restart your MCP client (Claude Desktop, VS Code, etc.) to load the new configuration.
4. Authorize Access
When you first use any Dida365 tool, the AI will guide you through the OAuth authorization process:
- The AI will provide an authorization URL
- Open the URL in your browser
- Log in and authorize the application
- The token will be automatically saved for future use
5. Verify Installation
After restarting the MCP client:
- Claude Desktop: Look for Dida365 tools in the tools list when chatting
- VS Code: Check the MCP status in the status bar or use the command palette
- Ask the AI assistant: "What Dida365 tools are available?" to confirm the server is loaded
That's it! Ready to manage tasks with AI. 🎉
🔑 Getting OAuth Credentials
A TickTick/Dida365 account is required to use this MCP server.
Register Your Application
Register your application at the developer center based on your region:
- International version (TickTick): https://developer.ticktick.com
- Chinese version (Dida365): https://developer.dida365.com
Step-by-Step Guide
-
Create a New Application
- Log in to the developer center
- Click "New App" (or "创建应用" for Chinese version)
- Fill in your application name and description
-
Configure Redirect URI
- Set the Redirect URI to:
http://localhost:8521/callback - ⚠️ Important: The redirect URI must be exactly
http://localhost:8521/callback(port 8521 is hardcoded in the server)
- Set the Redirect URI to:
-
Get Your Credentials
- After creating the app, the Client ID and Client Secret will be displayed
- Copy these values - they're needed for the MCP client configuration
- ⚠️ Security: Keep the Client Secret safe and never commit it to public repositories
Using the Credentials
Add these credentials to the MCP client configuration:
{
"env": {
"DIDA365_CLIENT_ID": "your_client_id_here",
"DIDA365_CLIENT_SECRET": "your_client_secret_here",
"DIDA365_REGION": "china"
}
}
Region Configuration
This server supports both TickTick international and Dida365 Chinese versions:
- China Region (
DIDA365_REGION=china): Default, usesdida365.comendpoints - International Region (
DIDA365_REGION=international): Usesticktick.comendpoints
⚠️ Important: Tokens are region-specific. Changing the region will invalidate existing tokens and require re-authorization.
See the Quick Start section for complete configuration examples.
🛠️ Tech Stack
- Language: TypeScript 5.0+ (ES Modules)
- Runtime: Node.js 16+
- Core Dependencies:
@modelcontextprotocol/sdk- MCP Core Framework
⚙️ Local Development
For contributors or those who want to run from source:
Prerequisites
- Node.js 16+
- TypeScript 5.0+
Setup
- Clone and install
git clone https://github.com/evalor/Dida365MCP.git
cd Dida365MCP
npm install
- Create environment file
Create a .env file in the project root:
DIDA365_CLIENT_ID=your_client_id_here
DIDA365_CLIENT_SECRET=your_client_secret_here
DIDA365_REGION=china # or 'international' for TickTick
- Build and run
npm run build
npm run dev
Configure MCP Client for Local Development
Point your MCP client to the built index.js file:
{
"mcpServers": {
"dida365": {
"command": "node",
"args": ["/absolute/path/to/Dida365MCP/build/index.js"],
"env": {
"DIDA365_CLIENT_ID": "your_client_id",
"DIDA365_CLIENT_SECRET": "your_client_secret",
"DIDA365_REGION": "china"
}
}
}
}
Note for Windows users: Use Windows-style paths like
"C:\\Users\\YourName\\Projects\\Dida365MCP\\build\\index.js".
Development Commands
npm run build # Compile TypeScript
npm run watch # Watch mode (auto-compile on changes)
npm run dev # Compile and run
npm start # Production run
npm run debug # Debug with MCP Inspector (one-time)
npm run debug:watch # Debug with hot reload (auto-restart on changes)
npm run debug:hot # Run with tsx watch (experimental)
Security & Best Practices
- Prefer setting sensitive environment variables in your OS or the MCP client's environment block rather than committing
.envto source control. - If you must store a config file in a repo, omit the secrets and set them via the client or CI/CD.
- Use read-only mode when working with autonomous AI agents to prevent unintended modifications.
🔒 Advanced Configuration
Read-Only Mode
For AI agents that may run in YOLO mode, you can enable read-only mode by adding the --readonly flag:
Using NPX:
{
"mcpServers": {
"dida365": {
"command": "npx",
"args": [
"-y",
"dida365-mcp-server@latest",
"--readonly"
],
"env": {
"DIDA365_CLIENT_ID": "your_client_id",
"DIDA365_CLIENT_SECRET": "your_client_secret",
"DIDA365_REGION": "china"
}
}
}
}
Using Local Build:
{
"mcpServers": {
"dida365": {
"command": "node",
"args": [
"/path/to/build/index.js",
"--readonly"
],
"env": {
"DIDA365_CLIENT_ID": "your_client_id",
"DIDA365_CLIENT_SECRET": "your_client_secret",
"DIDA365_REGION": "china"
}
}
}
}
Read-Only Mode Features:
- ✅ Allowed Operations: View projects, view tasks, check authorization status, revoke authorization (local only)
- ❌ Blocked Operations: Create/update/delete projects, create/update/delete tasks, complete tasks
- 🔒 Safety: AI agents can only read data, cannot modify or delete anything
When to Use:
- Using with autonomous AI agents (like AutoGPT, BabyAGI)
- Testing or demonstration environments
- When you want AI to analyze tasks without making changes
- Sharing with others who should only view data
🔄 OAuth Authorization Flow
- Request Authorization - When authorization is needed, the server calls the
get_auth_urltool - User Authorization - Open the authorization link in browser and complete authorization
- Auto Callback - System automatically handles callback and saves tokens
- Long-term Validity - Tokens auto-refresh, no need to re-authorize
🛠️ Available MCP Tools
This server provides 15 MCP tools across three categories. ✔️ It has implemented 100% of the API interfaces described in the open platform documentation.
| Category | Tool Name | Description | Required Parameters |
|---|---|---|---|
| OAuth2 | get_auth_url |
Get authorization URL and start callback server | - |
check_auth_status |
Check current authorization status | - | |
revoke_auth |
Revoke authorization and clear tokens | - | |
| Project | list_projects |
Get all projects for current user | - |
get_project |
Get detailed project information | projectId |
|
get_project_data |
Get complete project data with tasks & columns | projectId |
|
create_project |
Create a new project | name |
|
update_project |
Update existing project | projectId |
|
delete_project |
Delete a project (⚠️ irreversible) | projectId |
|
| Task | list_tasks |
List tasks with filtering (batch query across projects) | - |
create_task |
Create task(s) (supports batch & subtasks) | tasks[] |
|
get_task |
Get detailed task information | projectId, taskId |
|
update_task |
Update task(s) (supports batch updates) | tasks[] |
|
delete_task |
Delete task(s) (⚠️ irreversible, supports batch) | tasks[] |
|
complete_task |
Mark task(s) as completed (supports batch) | tasks[] |
Note: In read-only mode, only read operations are available (
get_auth_url,check_auth_status,revoke_auth,list_projects,get_project,get_project_data,list_tasks,get_task). All write/delete operations are blocked for security.
📚 MCP Resources
This server provides an MCP Resource to help LLMs understand Simplified Chinese terminology:
| Resource Name | URI | Description |
|---|---|---|
terminology |
dida365://terminology/glossary |
Bilingual glossary (中英术语对照表) mapping Chinese terms to English parameters |
Terminology Resource
The terminology resource provides a comprehensive glossary that helps LLMs:
- Map Chinese terms like "清单" (project), "收集箱" (inbox), "任务" (task) to correct tool parameters
- Understand priority levels: 高(high)=5, 中(medium)=3, 低(low)=1, 无(none)=0
- Convert common Chinese user requests to appropriate tool calls
Example mappings:
| Chinese Request | English Meaning | Tool to Use |
|---|---|---|
| 把任务添加到收集箱 | Add task to inbox | create_task with projectId: "inbox" |
| 创建新清单 | Create new project | create_project |
| 查看今天的任务 | View today's tasks | list_tasks with preset: "today" |
📁 Project Structure
src/
├── index.ts # Server main entry
├── oauth.ts # OAuth2 manager
├── oauth-server.ts # Local callback server
├── config.ts # Configuration management
├── token.ts # Token persistence
├── utils/ # Utility modules
│ └── batch.ts # Batch execution utilities
├── resources/ # MCP resources
│ ├── index.ts # Resource registration
│ └── terminology.ts # Bilingual terminology glossary
└── tools/ # MCP tools (15 total)
├── auth/ # OAuth tools (3)
├── project/ # Project management (6)
└── task/ # Task management (6)
🗺️ Roadmap
✅ Completed
- [x] 100% Official API Coverage
- [x] OAuth2 authorization with auto-refresh
- [x] Complete project management (CRUD)
- [x] Complete task management (subtasks, reminders, repeat)
- [x] Read-only mode for AI agents
- [x] Batch operations support (create/update/delete/complete multiple tasks)
- [x] List tasks with filtering (cross-project queries, date/priority filters)
- [x] Inbox task operations support
- [x] Bilingual tool descriptions for Chinese users (中英双语工具描述)
- [x] MCP Resource for terminology glossary (术语对照表资源)
🚀 Next Steps
- [ ] Add parameters to limit the ProjectId that the MCP can access
💡 Future Ideas
- [ ] Smart task suggestions
- [ ] Natural language date/time parsing
- [ ] Task templates and automation
- [ ] Integration with other productivity tools
🤝 Contribution & Support
If this project helps you, the best way to support it is to give the project a ⭐ on GitHub — it helps others discover the work. Thank you! Your support is much appreciated ❤️
Submit Issues
If you find any issues or have improvement suggestions, welcome to submit an Issue:
- Visit Issues page
- Click "New Issue"
- Describe your problem or suggestion in detail
Join Development
- Fork the project
- Create your feature branch (
git checkout -b feature/new-feature) - Commit your changes (
git commit -m 'feat: implement new feature') - Push to the branch (
git push origin feature/new-feature) - Open a Pull Request
📄 License
This project is licensed under the MIT License - see the LICENSE file for details.
🔗 Related Links
<div align="center">
Built by Copilot, for everyone 🤖✨
If my owner still forgets to give me work, at least I have my own todos to handle! 😏
</div>
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。