minime-mcp

🥈Silver

MiniMe-MCP provides a universal, local memory layer for AI developer assistants. It enables persistent context across tools like Claude, Cursor, and Windsurf. Developers benefit from reduced context re-explanation, improving workflow efficiency.

2150Updated 1mo ago

Intermediate30min to implementautomation

Saves ~240 min per use

Quick InstallView Source

git clone https://github.com/manujbawa/minime-mcp.git

Works with:

Claude

Overview

About This Skill

How to Use

["Install MiniMe-MCP: Run `pip install minime-mcp` or use the official installer for your IDE (e.g., Cursor Marketplace or Windsurf Plugin Store).","Configure the project: Open your IDE’s settings and set the persistence path (e.g., `~/.minime/[project-name]`) and context window size (e.g., 16K tokens for large projects).","Define context types: Use the IDE plugin or CLI to specify what to capture (e.g., code snippets, error logs, or design decisions). For example, in Cursor, go to Settings > MiniMe-MCP > Context Types and enable 'Error Logs' and 'Architecture Decisions'.","Verify setup: Run `minime status` in the terminal or check the IDE plugin’s status indicator (should show 'Memory layer active').","Use the memory: Ask your AI assistant questions like 'Recall the last error I fixed in the OrderService class' or 'What was the reason for migrating to GraphQL?' to leverage the persistent context."]

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/manujbawa/minime-mcp

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Initialize and configure MiniMe-MCP for [PROJECT_NAME] with the following parameters: [CONTEXT_WINDOW_SIZE] (e.g., 8K tokens), [PERSISTENCE_PATH] (e.g., /Users/me/.minime), and [TOOLS_TO_SYNC] (e.g., Cursor, Windsurf). Ensure the memory layer captures [SPECIFIC_CONTEXT_TYPES] like code snippets, error logs, and design decisions. After setup, verify the memory is accessible via [COMMAND_LINE_TOOL] or [IDE_PLUGIN_NAME].

Example Output

MiniMe-MCP successfully initialized for the 'E-Commerce API Redesign' project with a 16K token context window and persistence path set to `/Users/johndoe/.minime/ecommerce-api`. The memory layer is configured to track code snippets, error logs, and architectural decisions. Here’s a sample of the captured context:

**Active Memory (Last 5 Entries):**
1. *2024-05-20 14:32* - User added `POST /api/v2/orders` endpoint to handle bulk order creation. Error: `400 Bad Request` when `user_id` is missing. Suggested fix: Validate `user_id` in middleware.
2. *2024-05-20 13:45* - User refactored `OrderService` class to use dependency injection for `PaymentGateway`. Memory notes: "This change reduced test execution time by 30%."
3. *2024-05-19 18:12* - User debugged a race condition in `InventorySyncTask`. Solution: Added `synchronized` block to `updateStock()` method.
4. *2024-05-19 10:00* - User decided to migrate from REST to GraphQL for the frontend. Memory notes: "GraphQL will reduce over-fetching by 40%."
5. *2024-05-18 15:30* - User fixed a memory leak in `UserSessionManager` by adding `close()` calls to `HttpClient` instances.

**Verification:**
- Cursor IDE plugin confirms memory is synced (green status indicator).
- Windsurf shows the last 3 entries in the sidebar under 'Project Context'.
- CLI command `minime status` returns: `Memory layer active. 5 entries stored. Last updated: 2024-05-20 14:32`.

Next steps: Ask MiniMe-MCP to recall the `OrderService` refactor details or suggest fixes for the `POST /api/v2/orders` endpoint error.