dida365-agent-mcp 2.0.1

MCP server for AI Agents to manage Dida365 (TickTick) tasks and projects

dida365-agent-mcp

MCP server for AI agents to manage Dida365 / TickTick tasks and projects

中文文档

---

A Python MCP server built on FastMCP that connects AI agents (Claude Code, Cursor, Windsurf, etc.) to the Dida365 / TickTick Open API. Manage tasks and projects through natural language.

Supports both Dida365 (China) and TickTick (International) — switch with a single env var.

Features

• 40 tools — 19 V1 (official Open API) + 21 V2 (private API: tags, search, habits, folders, parent tasks)
• Dual transport — stdio for local clients; streamable-http for remote agents
• Dual platform — Dida365 and TickTick via DIDA365_REGION config
• One-click OAuth — script auto-opens browser, receives callback, saves token
• Docker ready — docker compose up to deploy
• Agent-friendly — structured JSON responses, clear tool annotations, actionable error messages

Deployment

Four options, ordered from simplest to most complex.

---

Option A — uvx (no install)

No git clone, no venv. Requires only uv.

1. Get API credentials

Create an app at the developer portal and copy the Client ID and Secret:

• Dida365: https://developer.dida365.com/manage
• TickTick: https://developer.ticktick.com/manage

Set Redirect URI to http://localhost:8000/oauth/callback.

2. Create .env and authorize

# Create config file
cat > .env << 'EOF'
DIDA365_REGION=china
DIDA365_CLIENT_ID=your_client_id
DIDA365_CLIENT_SECRET=your_client_secret
EOF

# Run OAuth — opens browser, saves token (~180 days)
uvx --from dida365-agent-mcp dida365-oauth

3. Connect your AI client

Claude Code

Edit ~/.claude/mcp.json:

{
  "mcpServers": {
    "dida365": {
      "command": "uvx",
      "args": ["dida365-agent-mcp"]
    }
  }
}

Cursor

Go to Settings > MCP > Add new global MCP server:

{
  "mcpServers": {
    "dida365": {
      "command": "uvx",
      "args": ["dida365-agent-mcp"]
    }
  }
}

---

Option B — pip install

pip install dida365-agent-mcp

Then follow the same steps as Option A, replacing uvx --from dida365-agent-mcp dida365-oauth with dida365-oauth and uvx dida365-agent-mcp with dida365-mcp.

---

Option C — git clone (source / development)

git clone https://github.com/linhai0872/dida365-agent-mcp.git
cd dida365-agent-mcp
uv sync
cp .env.example .env   # fill in CLIENT_ID and CLIENT_SECRET
uv run python scripts/oauth_flow.py

MCP client config:

{
  "mcpServers": {
    "dida365": {
      "command": "uv",
      "args": ["--directory", "/path/to/dida365-agent-mcp", "run", "dida365-mcp"]
    }
  }
}

---

Option D — Remote (Docker + HTTP)

For always-on or team deployments.

1. Get the access token first (run OAuth from Option A, B, or C)

2. Deploy

git clone https://github.com/linhai0872/dida365-agent-mcp.git
cd dida365-agent-mcp
cp .env.example .env
# Set DIDA365_REGION and DIDA365_ACCESS_TOKEN in .env
docker compose up -d

Connect to http://your-host:8000/mcp.

The container cannot open a browser, so set DIDA365_ACCESS_TOKEN directly instead of running the OAuth script inside Docker.

---

(Optional) Enable V2 tools — tags, search, habits, folders, parent tasks

Add one of the following to your .env (or MCP client env block):

# Method 1: Session token (manual, most secure, 30-day expiry)
# Browser DevTools → Application → Cookies → copy 't' value
DIDA365_V2_SESSION_TOKEN=your_token

# Method 2: Auto-login (convenient, does not support 2FA)
DIDA365_USERNAME=your_email
DIDA365_PASSWORD=your_password

Without V2 config, you still get 19 V1 tools. With V2, you get all 40.

---

Try It

Talk to your AI agent naturally:

• "List all my projects"
• "Create a high-priority task 'Review PR' in Work, due tomorrow"
• "What tasks did I complete this week?"
• "Move 'Design review' to the Archive project"
• "Search tasks containing 'meeting'" (V2)
• "List all my tags" (V2)
• "Did I check in my reading habit today?" (V2)

Tools

Task Operations

Tool	Description
dida365_create_task	Create a task with title, project, dates, priority, tags, reminders, recurrence
dida365_update_task	Update any task fields (partial update)
dida365_complete_task	Mark a task as completed
dida365_delete_task	Permanently delete a task
dida365_get_task	Get full details of a single task
dida365_move_task	Move a task between projects

Query Operations

Tool	Description
dida365_get_project_tasks	Get all uncompleted tasks in a project
dida365_filter_tasks	Filter by project, date range, priority, tags, status
dida365_get_completed_tasks	Get completed tasks within a time range

Project Operations

Tool	Description
dida365_list_projects	List all projects (call this first to get IDs)
dida365_get_project	Get project details
dida365_create_project	Create a project (list / kanban / timeline view)
dida365_update_project	Update project properties
dida365_delete_project	Permanently delete a project and all its tasks

Batch Operations

Tool	Description
dida365_get_task_by_id	Get a task by ID only (no project_id needed)
dida365_list_undone_tasks	List undone tasks by date range or project
dida365_batch_create_tasks	Batch create multiple tasks
dida365_batch_update_tasks	Batch update multiple tasks
dida365_batch_complete_tasks	Batch complete multiple tasks

V2 Tools (optional — requires V2 auth)

Tool	Description
dida365_search_tasks	Server-side full-text search with keyword, project, tag, status, date filters
dida365_list_tags	List all tags
dida365_create_tags	Batch create tags
dida365_update_tags	Batch update tags
dida365_delete_tags	Batch delete tags
dida365_delete_tag	Delete a single tag by name
dida365_set_task_parent	Set parent-child task relationship
dida365_unset_task_parent	Remove parent, make task top-level
dida365_pin_task	Pin or unpin a task
dida365_list_habits	List all habits
dida365_create_habit	Batch create habits
dida365_update_habit	Batch update habits
dida365_delete_habit	Batch delete habits
dida365_checkin_habit	Check in a habit for a date
dida365_undo_checkin	Undo a habit checkin
dida365_list_habit_checkins	List checkin records for a habit
dida365_list_habit_sections	List habit sections/groups
dida365_list_folders	List project folders
dida365_create_folder	Create a project folder
dida365_update_folder	Update a project folder
dida365_delete_folder	Delete a project folder

Configuration

Variable	Description	Default
DIDA365_REGION	china (dida365.com) or international (ticktick.com)	china
DIDA365_CLIENT_ID	OAuth Client ID	—
DIDA365_CLIENT_SECRET	OAuth Client Secret	—
DIDA365_ACCESS_TOKEN	Access token (set directly to skip OAuth)	—
DIDA365_REDIRECT_URI	OAuth callback URL	http://localhost:8000/oauth/callback
DIDA365_V2_SESSION_TOKEN	V2 session token (browser cookie t, 30-day expiry)	—
DIDA365_USERNAME	V2 auto-login email/phone (alternative to session token)	—
DIDA365_PASSWORD	V2 auto-login password (does not support 2FA)	—
TRANSPORT	stdio, streamable-http, or sse (legacy)	stdio
HOST	Bind address (http / sse only)	0.0.0.0
PORT	Port (http / sse only)	8000

Token Lifecycle

Item	Detail
Validity	~180 days
Auto-refresh	Not supported (API limitation)
Expiry detection	Built-in, warns 24h before expiry
Renewal	Re-run dida365-oauth (or uvx --from dida365-agent-mcp dida365-oauth)
Storage	~/.dida365-agent-mcp/token.json (auto-loaded)

Development

uv sync                                    # Install dependencies
uv run pytest tests/ -v                    # Run tests
uv run ruff check src/ tests/ scripts/     # Lint
uv run ruff format src/ tests/ scripts/    # Format
uv run pyright src/                        # Type check
uv run dida365-mcp                         # Run (stdio)
TRANSPORT=streamable-http uv run dida365-mcp   # Run (Streamable HTTP)

Project Structure

dida365-agent-mcp/
├── src/dida365_agent_mcp/
│   ├── server.py        # FastMCP server + 19 V1 tool definitions + MCP resource
│   ├── server_v2.py     # 21 V2 tool definitions (tags, search, habits, folders)
│   ├── client.py        # V1 async API client (httpx + OAuth)
│   ├── client_v2.py     # V2 async API client (httpx + session cookie)
│   ├── auth.py          # OAuth2 flow + token management
│   ├── models.py        # Pydantic data models (V1 + V2)
│   └── config.py        # Region-aware configuration
├── scripts/
│   └── oauth_flow.py    # One-click OAuth script
├── tests/               # 62 unit tests (respx)
├── Dockerfile           # Multi-stage build
└── docker-compose.yml

License

MIT

Acknowledgments

• Dida365 / TickTick Open API
• FastMCP — Python MCP framework
• Model Context Protocol