Routine

Routine is an MCP server that connects AI assistants like Claude Desktop to the Routine desktop productivity application. Designed for professionals, project managers, and knowledge workers who rely on Routine to organize their workday, this integration exposes core workspace functionality directly to language models via the Model Context Protocol. By linking to a running instance of the Routine application, the server enables users to review their daily agenda, inspect upcoming events, and track pending to-do items using conversational natural language prompts. Users can also instruct their AI client to capture and create new tasks immediately, streamlining task capture during drafting or research sessions without switching windows. Furthermore, the server allows assistants to locate and open specific pages, documents, and notes stored in the Routine app. The server communicates through standard input and output channels, facilitating local interaction between AI clients and local application data. By maintaining access to calendar items, notes, and task lists, the server helps users coordinate their workflow, prepare for upcoming meetings, and keep their personal knowledge base synchronized without leaving their preferred AI conversational interface.

Category: Communication & Messaging

Tags: calendar, notes, productivity, tasks

Visit Routine

How to install and configure Routine

1. Ensure the Routine desktop application is installed and actively running on your computer. 2. Open your Claude Desktop configuration file (claude_desktop_config.json). 3. Add the Routine MCP server entry under the mcpServers object: json { "mcpServers": { "routine": { "command": "npx", "args": ["routine-mcp-server"] } } } 4. Save the configuration file and restart Claude Desktop. Alternatively, for local development, clone the repository, run yarn to install dependencies, run yarn build, and point your MCP configuration command to node with the absolute path to ./dist/index.js.

What you can do with Routine

• Checking daily schedules: Request your AI assistant to retrieve upcoming calendar events and pending agenda items scheduled for today in Routine. - Creating new to-do tasks: Convert conversational action items directly into actionable Routine tasks with natural language prompts during planning sessions. - Opening specific notes: Prompt your assistant to locate and launch a dedicated Routine note or documentation page inside the desktop client. - Reviewing upcoming deadlines: Ask the assistant to query your Routine calendar and task manager to summarize near-term deadlines and obligations.

Key facts

• https://github.com/routineco/mcp-server
• Communication & Messaging, Notes & Knowledge Management, Project & Task Management
• calendar, notes, productivity, tasks

Part of MCP Servers

How do I install Routine MCP server?

You can run the Routine MCP server using npx without manual installation. Ensure the Routine desktop application is open and running on your machine, then add the routine configuration entry with command npx and argument routine-mcp-server to your client configuration file, such as claude_desktop_config.json. Restart your client to finalize the setup.

What can Routine MCP server do?

The server lets supported AI assistants interact with your Routine workspace. You can ask your assistant to check your daily schedule, retrieve upcoming calendar events, create new tasks from chat discussions, and open specific notes or documentation pages directly in the Routine desktop application.

Does Routine need to be running for the MCP server to work?

Yes, the Routine desktop application must be active and running on your local machine. The MCP server communicates directly with your local Routine app instance to read your schedule, create new tasks, and open pages or documents inside the user interface.

Which MCP clients work with Routine?

Any client supporting the Model Context Protocol over standard input and output can connect to this server. Routine provides direct setup instructions for Claude Desktop, but developers can also test and integrate it with other compliant MCP hosts by executing the npx routine-mcp-server command.