Claude Todo MCP Server
# Claude Todo MCP Server
[](https://pypi.org/project/claude-todo-emulator/)
[](https://pepy.tech/projects/claude-todo-emulator)
**Add persistent task management to any AI coding assistant in Cursor, Windsurf, and other IDEs.**
This MCP server replicates Claude Code's todo system, giving your AI assistant the ability to:
- Track complex multi-step tasks across sessions
- Break down large features into manageable pieces
- Remember progress when you switch between projects
- Enforce single in-progress task focus
## Quick Start
**No installation required!** Just add this to your MCP configuration:
### Cursor IDE
Create `.cursor/mcp.json` in your project root:
```json
{
"mcpServers": {
"task-manager": {
"command": "uvx",
"args": ["claude-todo-emulator"]
}
}
}
```
### Codex CLI
Create `config.toml` at your project root:
```toml
[mcp_servers.task-manager]
command = "uvx"
args = ["claude-todo-emulator"]
```
### Other IDEs
Add the JSON configuration above to your MCP settings file.
**Restart your IDE** and try: _"Create a todo list for adding user authentication"_
The system automatically downloads the package and sets up everything on first use!
## What You Get
Once configured, your AI assistant will automatically:
**Create todos for complex tasks:**
```
You: "Add dark mode support with user preferences"
AI: "I'll create a todo list for this multi-step feature:
- ✅ Add theme context and state management
- 🔄 Create dark mode CSS variables and styles
- ⏳ Build settings UI for theme switching
- ⏳ Add persistence to localStorage
- ⏳ Update existing components for theme support"
```
**Track progress visually:**
```
You: "What's the status on the authentication system?"
AI: "Current authentication tasks:
- ✅ Set up JWT middleware
- ✅ Create login/signup endpoints
- 🔄 Building user profile management (in progress)
- ⏳ Add password reset flow
- ⏳ Implement role-based permissions"
```
**Maintain focus:**
```
You: "Let's also add email notifications"
AI: "I'll add that to the todo list, but let me finish the current user profile task first to maintain focus."
```
### API Reference
**TodoRead**
- Parameters: None
- Returns: `{todos: Todo[]}`
**TodoWrite**
- Parameters: `{todos: Todo[]}`
- Returns: `{success: boolean, count: number}`
**Todo Schema**
```typescript
{
id: string;
content: string;
status: "pending" | "in_progress" | "completed";
priority: "high" | "medium" | "low";
created_at: string;
updated_at: string;
metadata?: object;
}
```
## Contributing
1. Fork the repository
2. Create a feature branch
3. Add tests for new functionality
4. Submit a pull request
## License
MIT License - see LICENSE file for details.
## Related Projects
- [Model Context Protocol](https://modelcontextprotocol.io/)
- [FastMCP](https://gofastmcp.com/)
- [Claude Code](https://www.anthropic.com/claude-code)
TDQS
Scored across 2 tools
The two tools have clearly distinct purposes: TodoRead is for retrieving the task list, while TodoWrite is for replacing it entirely. There is no overlap or ambiguity between these operations.
Both tools follow a consistent TodoVerb naming pattern (TodoRead and TodoWrite), using the same prefix and clear action verbs. This makes their functions immediately understandable and predictable.
With only two tools, the server is severely limited for a todo management domain. It lacks essential operations like creating, updating, or deleting individual tasks, making it impractical for typical todo workflows.
The tool surface is severely incomplete. It only supports reading and full replacement of the task list, missing critical CRUD operations such as add_todo, update_todo, delete_todo, or mark_complete. This will cause frequent agent failures in managing tasks.