Roo Code Memory Bank MCP Server

🚀 Roo Code Memory Bank MCP Server

This project implements the core functionality of the Roo Code Memory Bank system as a Model Context Protocol (MCP) server. It allows AI assistants to maintain project context across sessions by interacting with a file-based memory bank.

🚀 Quick Start

This MCP server enables AI assistants to interact with a file-based memory bank, maintaining project context across sessions.

✨ Features

This MCP server offers the following tools:

• initialize_memory_bank: Creates the memory-bank/ directory and generates standard .md files (productContext.md, activeContext.md, progress.md, decisionLog.md, systemPatterns.md), populating them with initial templates.

  • Input: (Optional) { "project_brief_content": string }
  • Output: { "status": "success" | "error", "messages": Array<string> }

• check_memory_bank_status: Checks if the memory-bank/ directory exists and returns its status.

  • Input: None
  • Output: { "exists": boolean, "message": string }

• read_memory_bank_file: Reads and returns the content of a specified .md file.

  • Input: { "file_name": string }
  • Output: { "content": string, "error": null | string }

• append_memory_bank_entry: Appends specified content to a specified .md file.

  • Input: { "file_name": string, "entry_content": string }
  • Output: { "status": "success" | "error", "message": string }

• initialize_memory_bank: Initializes a new memory bank, including generating all necessary initial files and directory structures.

  • Input: None
  • Output: { "status": "success" | "error", "message": string }

📦 Installation

To use this MCP server, ensure the following:

1. You have Node.js and npm installed.
2. You are familiar with basic command-line operations and file management.

Follow these steps to install and run the MCP server:

1. Clone this repository to your local machine.
2. Open a terminal or command prompt and navigate to the repository's root directory.
3. Run npm install to install all dependencies.
4. Start the server: node index.js or use npm start.

📚 Documentation

Configuration

For the Cline MCP client, configure as follows:

Add the following to your project's config.json file:

{
  "mcp_servers": {
    "roo_code_memory_bank": {
      "url": "http://localhost:3000",
      "auth_token": "your_auth_token_here"
    }
  }
}

Running the Server

The server runs locally on port 3000 by default.

Usage Instructions

• Use check_memory_bank_status at the start of any task or important subtask.
• If the memory bank exists (exists: true), use read_memory_bank_file to read relevant files (e.g., productContext.md, activeContext.md) to load the current project context.
• Integrate this loaded context into your planning and execution.
• When making important decisions, updating progress, or changing the architecture, use append_memory_bank_entry to record information in appropriate files (e.g., decisionLog.md, progress.md) to ensure context persistence.
• If the memory bank does not exist, consider using initialize_memory_bank in your project.

Custom Instructions

Set these instructions in Roo:

You must use MCP where needed.

I have a specific MCP process:

1. Use context7 to find any relevant document fragments. You'll need to do this for each subtask. Ensure all relevant knowledge is passed to any relevant subtasks, and always use context7 for research if you're unsure about something.
2. Use the google maps mcp to search for <niche> + <county>. This will help us find basic business information needed to complete tasks.
3. Use the brave search mcp to find URLs for scraping.
4. Use the fetch mcp with fetch_txt and fetch_markdown to find text and images on pages and convert them into in-depth JSON files.

Leverage the roo-code-memory-bank-mcp server to maintain project context:

• Use check_memory_bank_status at the start of tasks or important subtasks.
• If the memory bank exists (exists: true), use read_memory_bank_file to read relevant files (e.g., productContext.md, activeContext.md) to load the current project context.
• Integrate this loaded context into your planning and execution.
• When making important decisions, updating progress, or changing the architecture, use append_memory_bank_entry to record information in appropriate files (e.g., decisionLog.md, progress.md) to ensure context persistence.
• If the memory bank does not exist, consider using initialize_memory_bank in your project.