minimi

Overview

Minimi is a macOS menu bar application that acts as an ambient, on-device memory layer for Claude. It quietly captures and indexes activity across browser tabs, documents, calls, and chat threads to build a private personal memory store. Through a Model Context Protocol (MCP) connector, it feeds live context directly to Claude so users can query past decisions without manual context reloading. The system helps knowledge workers eliminate repetitive re-briefing while maintaining strict local data ownership.

What it does

• Captures digital activity across macOS apps, documents, meetings, and browser tabs continuously in the background.
• Indexes local vector embeddings privately on-device without exposing raw content to cloud servers.
• Connects seamlessly to Claude via a Model Context Protocol (MCP) custom connector interface.
• Surfaces closed loops, action items, and cross-app team insights through natural language queries.

Who it's for

• Software Engineers
• Product Managers
• Founders
• Knowledge Workers

Why it works

• Local-first vector storage ensures high privacy and keeps user memory data securely on their Mac.
• Model Context Protocol (MCP) native integration allows direct data streaming into Claude without copy-pasting.
• Install-and-forget workflow runs completely in the background without requiring manual context-saving steps.
• Benchmarked high long-term recall accuracy outperforms traditional generic RAG and search pipelines.

Founder story

Minimi was developed by Shram Intelligence, Inc., based in Lewes, Delaware, to solve the friction of starting from scratch every time users open an AI assistant. Built by a team focused on workflow intelligence, the tool provides an ambient memory layer that tracks daily professional context across macOS environments.

Alternatives

• Rewind AI
• Limitless
• Mem
• Superwhisper

Comparison overview

• Operates as a dedicated native macOS menu bar app with deep Model Context Protocol (MCP) integration for Claude , Stores embeddings locally on-device for enhanced privacy compared to standard cloud-synced memory tools , Focuses specifically on capturing ambient work context across disparate applications without manual logging