Memory Bank

Provides a guided MCP server to initialize, read, and update a structured memory bank for AI-assisted project context management.
  • javascript

17

GitHub Stars

javascript

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "hoppo-chan-memory-bank-mcp": {
      "command": "npx",
      "args": [
        "@neko0721/memory-bank-mcp"
      ]
    }
  }
}

Memory Bank MCP helps AI assistants persist and organize project context across sessions using a structured set of coremarkdown files. It guides you through initializing, reading, and updating memory content so your AI workflows stay aware of goals, decisions, progress, and patterns across your projects.

How to use

You interact with Memory Bank MCP through an MCP client to initialize, read, and update the memory context for your project. Your AI assistant can check the current project state, set up memory files when starting a new project, and apply guided updates after tasks or conversations so the memory remains accurate and useful.

How to install

Prerequisites you need to prepare before using Memory Bank MCP:

  • Node.js (LTS version) and npm are installed on your system
  • A project directory where you want to manage memory context
  • A code editor to edit or create the memory files

Step-by-step setup to enable Memory Bank MCP for your project:

# No installation is required beyond ensuring Node.js is available.
# Create or navigate to your project directory
mkdir -p path/to/your/project
cd path/to/your/project

# Create the MCP configuration to enable Memory Bank
# This uses the published MCP package via npx
# File: mcp_config.json

Core usage steps

  1. Initialize Memory Bank when you start a new project. This creates the memory-bank directory and the core context files.

  2. Read Memory Bank to understand current project state and context. This shows goals, decisions, progress, and patterns.

  3. Update Memory Bank to get guided instructions for updating specific files as the project evolves.

Configuration overview

The MCP configuration demonstrates how Memory Bank is wired for operation. It uses a stdio (local) MCP server that runs via npx and loads the memory-bank MCP package.

{
  "mcpServers": {
    "memory_bank": {
      "command": "npx",
      "args": ["@neko0721/memory-bank-mcp"]
    }
  }
}

Usage guidelines

For best results, start each session by ensuring the memory-bank directory exists and then read the current memory content with the read command. When your work is complete or you’ve completed a planning or coding task, update Memory Bank so the memory reflects the latest context and decisions.

Update triggers

Trigger updates when major changes occur, such as architecture decisions, feature completions, bug fixes, refactoring, important decisions, or progress updates. These events should prompt you to refresh the memory context accordingly.

Tool reference

The Memory Bank MCP tools help you initialize, inspect, and update the memory context.

Integration tips

Cursor Setup and Windsurf Setup tips are provided to ensure you consistently check for and initialize Memory Bank at session start, read context, and apply updates after tasks or conversations.

Best practices and notes

  • Memory Bank maintains five core files to organize your project context: productContext.md, activeContext.md, progress.md, decisionLog.md, and systemPatterns.md. These files help you capture goals, current work, progress, decisions, and patterns in a structured way.

Available tools

init-memory-bank

Initializes Memory Bank with all core files and guidance to start tracking project context.

get-memory-bank-info

Reads and returns all Memory Bank content to understand current project state.

update-memory-bank

Provides guided updates for Memory Bank files based on the described change type and description.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational