Chisle
The coding-agent plugin that bills 52% of a bare model across 20 live tasks. Terse persona + tool-output compressor + context diet.
翻訳待ち
Things.app MCP Server
翻訳待ち
開発ツールPython · ShellGitHub @hald自分でインストール所有者未確認
Paste into Claude or ChatGPT for installation instructions suited to your environment.
Have you tried it? One tap is enough. No sign-in needed.
The maker's introduction consists of the line above. Read the repository README for more details. Repository README ↗
Loading the full README.
Read the full README ↗This Model Context Protocol (MCP) server lets you use ChatGPT, Codex, Claude Desktop, and other MCP clients to interact with your task management data in Things 3 from Cultured Code. You can ask your assistant to create tasks, analyze projects, help manage priorities, and more.
This server leverages the Things.py library and the Things URL Scheme.
If you find this project helpful, consider supporting its development:
brew install uvThings MCP is published on PyPI and can be run directly with uvx:
uvx things-mcp
Configure your MCP client to use uvx with things-mcp as the argument.
.mcpb file{
"mcpServers": {
"things": {
"command": "uvx",
"args": ["things-mcp"]
}
}
}
claude mcp add-json things '{"command":"uvx","args":["things-mcp"]}'
To make it available globally (across all projects), add -s user:
claude mcp add-json -s user things '{"command":"uvx","args":["things-mcp"]}'
The ChatGPT desktop app, Codex CLI, and Codex IDE extension share the same local MCP configuration. The quickest way to add Things MCP is with the Codex CLI:
codex mcp add things -- uvx things-mcp
Confirm that the server is configured:
codex mcp list
You can also configure it in the ChatGPT desktop app:
things as the name, choose STDIO, and enter the result of which uvx as the command to launch (should look similar to /opt/homebrew/bin/uvx)things-mcp under arguments[!IMPORTANT] When using the ChatGPT desktop client, you must use Work mode to access Things MCP. Select ChatGPT from the top-left menu, then choose Work from the toggle at the top of the page. Regular Chat mode does not use locally configured MCP tools. You can also select the separate Codex view, which shares the same MCP configuration.
In ChatGPT Work or Codex, type /mcp to check that the Things server is connected, then try asking, "What's in my Things inbox?"
This local setup does not work in ChatGPT on the web because the web client does not read local Codex MCP configuration. See OpenAI's MCP documentation and ChatGPT Work and Codex for current availability and interface details.
After installation:
/mcp and confirm that the Things server is connectedget-inbox - Get todos from Inboxget-today - Get todos due todayget-upcoming - Get upcoming todosget-anytime - Get todos from Anytime listget-someday - Get todos from Someday list, including tasks in Someday projectsget-logbook - Get completed and canceled todos, optionally for one projectget-trash - Get trashed todosget-todos - Get todos, optionally filtered by projectget-projects - Get all projectsget-areas - Get all areasget-tags - Get all tagsget-tagged-items - Get items with a specific tagget-tag-usage - Report how many items use each tag, sorted by usage; flag unused tags for cleanupsearch-todos - Simple search by title/notessearch-advanced - Advanced search with multiple filtersThis is the beginning of the README. Read the rest at the original. Read the full README ↗