Apple-Shortcuts
Extend Claude's capabilities using MCP and MacOS Shortcuts
claude mcp add --transport stdio as2811-project-apple-shortcuts-mcp uv --directory ABSOLUTE/PATH/TO/THIS/DIRECTORY run main.py \ --env GEMINI_API_KEY="your_gemini_api_key"
How to use
This MCP server exposes a macOS automation workflow through the Claude/MCP integration by leveraging the built-in Shortcuts on macOS. The Shortcuts-MCP tool chain enables Claude to trigger a set of Mac Shortcuts that perform real-world tasks: extracting ingredients from a recipe using Gemini Video Understanding API, adding those ingredients to a Reminders list called Groceries, summarizing a chat thread and saving the summary as a note, and creating a calendar event. To use it, ensure you have the required Shortcuts installed on your Mac and that Claude has access to run the CLI-based uv server which runs the main.py script. When you provide Claude with a URL or a description, the MCP can chain these tools (e.g., first fetch ingredients, then add them to Groceries, and finally create a calendar event) by invoking the corresponding Shortcuts via the Python/uv runtime.
Usage scenarios include: extracting ingredients from a recipe URL and pushing them to a shopping list, summarizing a conversation and saving the result as a note, and creating calendar events from contextual input. You can daisychian multiple actions by giving Claude a sequence like “get ingredients from this recipe, add to Groceries, then create a calendar event for tomorrow at 3pm.” Ensure your Gemini API key is configured in the environment so the Gemini-based ingredient extraction works properly.
How to install
Prerequisites:
- macOS with Shortcuts available
- Python 3.x installed
- uv (uvx) installed globally on your system (brew install uv) if using the provided instructions
- Access to Gemini API key for ingredient extraction
Step-by-step installation:
-
Clone the repository: git clone <repository-url> cd <repository-directory>
-
Install uv and dependencies:
- On macOS, install uv if not already installed: brew install uv
- Install Python dependencies if any are required by main.py (e.g., google-genai) using pip: uv sync uv pip install google-genai
-
Prepare the configuration: Copy the Claude Config JSON below and replace ABSOLUTE/PATH/TO/THIS/DIRECTORY with the actual absolute path to this repo on your machine. Add your Gemini API key to GEMINI_API_KEY in the env section.
{ "mcpServers": { "Shortcuts-MCP": { "command": "uv", "args": [ "--directory", "ABSOLUTE/PATH/TO/THIS/DIRECTORY", "run", "main.py" ], "env": { "GEMINI_API_KEY": "your_gemini_api_key" } } } }
-
Start the MCP server:
- Run the server using the mcp_config you prepared. For example: uv --directory /absolute/path/to/this/directory run main.py
-
Open Claude and ensure the tools appear in the toolbox. You should see the Shortcuts-MCP tool configured and ready to use.
Additional notes
Tips and notes:
- Ensure you have created and named the Reminders list called Groceries on your Mac before using the Shortcuts that push items there.
- The Gemini API key is required for ingredient extraction; keep it secure and rotate as needed.
- The provided configuration uses the uv runner to execute main.py from a specific directory; ensure the path is correct and accessible.
- If Claude cannot see the tools, re-run uv sync and verify that main.py is present in the specified directory and that dependencies like google-genai are installed.
- You can chain multiple actions by instructing Claude to execute subsequent tasks after the previous one completes (e.g., extract ingredients, then add to Groceries, then create a calendar event).
- If you change the path or environment, update the mcp_config accordingly and restart the MCP service.
Related MCP Servers
web-eval-agent
An MCP server that autonomously evaluates web applications.
zotero
Model Context Protocol (MCP) server for the Zotero API, in Python
google_ads_mcp
The Google Ads MCP Server is an implementation of the Model Context Protocol (MCP) that enables Large Language Models (LLMs), such as Gemini, to interact directly with the Google Ads API.
fhir
FHIR MCP Server – helping you expose any FHIR Server or API as a MCP Server.
apple-books
Apple Books MCP Server
mcp-oi-wiki
🌟 Wiki of OI / ICPC for LLMs. (面向大模型的某大型游戏线上攻略,内含炫酷算术魔法)