Things MCP Server

Things MCP Server – Bridging Claude Desktop and Things 3

The Things MCP server fills a critical gap for developers who want to bring the power of AI into their task‑management workflow. It exposes the full data model of Things 3—tasks, projects, areas, tags, and checklists—to Claude Desktop via the Model Context Protocol. Instead of manually typing commands or navigating the Things UI, a developer can ask Claude to create a new task, move items between lists, or analyze productivity patterns—all through natural language. This reduces friction and unlocks a new level of automation for teams that already rely on Things as their primary productivity hub.

At its core, the server is a thin wrapper around Things.py and the Things URL Scheme. It translates MCP tool calls into API requests or URL‑scheme invocations, then returns structured JSON back to Claude. The implementation is engineered for reliability: exponential‑backoff retries, a circuit breaker that stops cascading failures when Things becomes unresponsive, and a dead‑letter queue to capture operations that ultimately fail. Intelligent caching of recent queries keeps latency low, while rate‑limiting protects the Things application from being overwhelmed by rapid requests.

Key capabilities include:

• Full list access (Inbox, Today, Upcoming, Custom lists) and manipulation of items within them.
• Project and area hierarchy management, enabling creation, renaming, or re‑assignment of nested structures.
• Tag operations for quick filtering and categorization.
• Advanced search with support for Boolean queries, due dates, and custom fields.
• Recent item tracking, allowing Claude to reference the last few tasks a user interacted with.
• Detailed checklist handling so sub‑tasks can be added, completed, or reordered.

Real‑world scenarios benefit from this integration: a developer can ask Claude to “create a task for the upcoming sprint review in Project X with due date tomorrow and tag it ‘Sprint’,” or “show me all overdue tasks in the Inbox.” Product managers can have Claude generate a prioritized list of backlog items based on stakeholder tags. Teams using the GTD methodology can let Claude suggest next actions or review projects, all while keeping everything in sync with Things 3.

For AI workflows, the server acts as a first‑class MCP provider. Claude Desktop can discover the available tools automatically and invoke them with minimal context, enabling seamless chaining of actions. Because the server handles authentication tokens securely (via a configuration script or environment variable) and respects Things’ URL‑scheme limits, developers can focus on building higher‑level conversational logic rather than plumbing the connection. The result is a robust, production‑ready bridge that turns natural language into precise task‑management actions—an indispensable asset for any developer looking to embed AI directly into their daily productivity stack.