minimi 2.0

Product Introduction

1. Definition: minimi 2.0 is a personal intelligence and context-aware AI assistant application for macOS. It operates as a local-first, on-device agent designed to automate personal task management by understanding user context.
2. Core Value Proposition: minimi 2.0 exists to eliminate the cognitive load of manual task tracking and context-switching. Its primary function is to automatically identify "open loops" (unresolved tasks or commitments) from a user's digital activity and resolve them, acting as an automated personal productivity manager.

Main Features

1. Memory MCP (Model Context Protocol): This is the core technology that captures and structures personal context. It works by continuously analyzing on-screen content via macOS's native accessibility APIs (not screenshots) to build a private, local knowledge graph of the user's work, conversations, and commitments. This memory can be exported as an MCP server to provide context to other AI tools like Claude or ChatGPT.
2. Open Loops Inbox & Automation: The AI scans the user's captured memory to detect actionable items, deadlines, and promises made across apps like Gmail, WhatsApp, and Calendar. It then automatically creates reminders, updates tasks, or triggers actions to "close" these loops without requiring manual user input or complex integrations.
3. Local-First, Privacy-Centric Architecture: All user data processing and storage occur locally on the user's Mac. The software does not send personal context or memory data to cloud servers, ensuring privacy. This architecture means the AI's intelligence is personalized to the individual device without external data transmission.

Problems Solved

1. Pain Point: "Context switching fatigue" and "inbox zero anxiety." Professionals lose significant productivity manually tracking action items scattered across different communication and productivity apps, leading to missed commitments and mental overhead.
2. Target Audience: Knowledge workers who context-switch frequently (e.g., project managers, executives, consultants), remote teams coordinating across multiple apps, and privacy-conscious individuals seeking AI assistance without data sharing to the cloud.
3. Use Cases: Automatically creating a to-do item from a promise made in a WhatsApp chat; adding a meeting follow-up task after a Calendar event ends; surfacing a commitment from an old email thread before a deadline; providing a unified memory context for external LLMs to answer personal questions accurately.

Unique Advantages

1. Differentiation: Unlike traditional task managers (Todoist, Asana) that require manual entry, or cloud-based AI assistants that process data externally, minimi 2.0 is fully automatic and local. It differs from other automation tools (Zapier, IFTTT) by using AI to identify what needs to be automated, rather than requiring user-defined rules.
2. Key Innovation: The integration of a local, continuously updated Memory MCP with an autonomous agent capable of taking action. The product's metaphor of "AI cats" (Cotton for memory, Melody for action) simplifies the complex technology of persistent memory and autonomous task completion into a user-friendly model.

Frequently Asked Questions (FAQ)

1. How does minimi 2.0 work without integrations? minimi 2.0 uses macOS's system-level accessibility features to read on-screen text and context from any application, allowing it to work with apps like WhatsApp, Gmail, and Calendar without requiring official APIs or user-built connections.
2. Is my data private with minimi's AI? Yes, minimi 2.0 employs a strict local-first architecture. All personal context and memory data are stored exclusively on your Mac device. The company states it does not store any user data on its servers or use it for AI model training.
3. What is an "open loop" in minimi? An "open loop" is any incomplete task or unresolved commitment that minimi's AI identifies from your digital communications and activity, such as an unanswered question, a promise to send a file, or an upcoming deadline without a prepared action.
4. Can I use minimi's memory with ChatGPT or Claude? Yes, through the minimi MCP (Model Context Protocol). You can connect your local minimi memory as a context source to compatible AI platforms, allowing LLMs like Claude or ChatGPT to access your personal context to provide more relevant and accurate answers.
5. What happens if minimi makes a mistake or misinterprets a task? As an autonomous agent, users should supervise its actions. The product is designed to learn and refine its understanding over time based on user corrections and continued context gathering, operating as a semi-autonomous assistant that improves with use.