jlumbroso/passage-of-time-mcp

What it solves

LLMs often struggle with mental math regarding time and lack a real-time sense of the current date and hour. This project provides a set of tools that allow models to accurately calculate time differences, understand the human context of a timestamp (such as whether it is a weekend or business hours), and maintain temporal awareness during a conversation.

How it works

It is implemented as a Model Context Protocol (MCP) server. When connected to an MCP-compatible client (like Claude.ai), the LLM can call specific functions to handle time-related tasks. The server provides tools for fetching the current date and time, calculating durations between timestamps, parsing various time formats, and adding or subtracting time durations.

Who it’s for

Developers and AI enthusiasts who want to give their LLM agents more precise temporal reasoning capabilities and a better understanding of human rhythms and work patterns.

Highlights

• Temporal Context: Goes beyond simple timestamps by identifying if a time is "dinner time," "leisure time," or "business hours."
• Human-Readable Output: Converts raw seconds into intuitive descriptions like "2 hours ago" or "3 hours, 10 minutes, 1 second."
• Collaborative Design: The toolset was developed through a partnership between a human and an LLM to identify what the model actually needed to perceive time.
• Timezone Support: Supports all standard timezone names to ensure global accuracy.