dida365-agent

Let AI agents manage Dida365 / TickTick tasks, projects, tags, and habits through natural language. Three form factors: a lightweight CLI, a ready-to-install Agent Skill, and a standard MCP Server.

No install required — run it directly with uvx. Supports both Dida365 (China) and TickTick (International), switchable with one env var.

Features

Related MCP server: TickTick/Dida365 MCP Server

Quick Start

Option 1: Use the CLI directly

No install needed — run with uvx (requires uv). First prepare credentials per the configuration guide and write them to .env:

# Browser OAuth (saves token, ~180 days)
uvx dida365-agent dida auth login

# List all projects
uvx dida365-agent dida project list

# Create a high-priority task due tomorrow
uvx dida365-agent dida task create --title "Review PR" --project <projectId> \
  --priority 5 --due-date "2026-05-30T18:00:00+0800"

# Full-text search (V2)
uvx dida365-agent dida search "meeting"

After installing locally (uv tool install dida365-agent), use the shorter dida command.

Option 2: Use as an Agent Skill

Install the Skill in any Skill-compatible AI tool (Claude Code, Cursor, etc.) and drive it with natural language:

# Install the Skill
npx skills add linhai0872/dida365-agent

Then just describe what you need in the AI chat:

Tidy up my unfinished tasks for today, list the high-priority ones, and move the overdue ones to the "Later" project

The Skill recognizes intent, fills in missing details, and assembles dida commands automatically — no manual parameters required.

Option 3: As an MCP Server

Exposes 44 tools as a standard MCP Server for Claude Code, Cursor, Windsurf, etc. See MCP Server integration.

Documentation

Development

uv sync                              # Install dependencies
uv run python -m pytest tests/       # Run tests
uv run ruff check src/ tests/        # Lint
uv run dida --help                   # Run the CLI locally

License

MIT

Acknowledgments

• Dida365 / TickTick Open API

• FastMCP · Typer · Model Context Protocol