Roo Code Memory Bank

🚀 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 and share project context through persistent storage.

✨ Features

Initialize the Memory Bank

• Create the memory-bank directory.
• Generate initial template files: productContext.md, activeContext.md, decisionLog.md, progress.md, and architecturalDecisions.md.

Check the Memory Bank Status

• Confirm whether the memory bank exists (returns a boolean value).

Read File Content

• Load and parse the specified .md file to obtain project context information.

Append Entry Records

• Append new entries or update existing content in relevant files to ensure the recording of important decisions, progress, and architectural changes.

💻 Usage Examples

Basic Usage

1. Initialize the Memory Bank
   • Use initialize_memory_bank to create the basic structure at the start of a new project.
2. Check the Memory Bank Status
   • Call check_memory_bank_status to confirm the existence of the context before starting a task or an important subtask.
3. Load Context Information
   • Use read_memory_bank_file to load the content of relevant files, such as productContext.md and activeContext.md, to obtain the current project status.
4. Record Important Changes
   • When making decisions or updating progress, use append_memory_bank_entry to add records to decisionLog.md or progress.md.

Advanced Usage

Custom Workflow

1. Check the Memory Bank at Startup
   • Use check_memory_bank_status to confirm the existence of the context.
2. Load the Current Context
   • Call read_memory_bank_file to load the content of relevant files, ensuring that the AI assistant gets the latest information.
3. Record Important Events
   • Use append_memory_bank_entry to record decisions, progress, or architectural changes.

MCP Tool Integration

• Context7: Find and integrate required documents.
• Google Maps: Search for enterprises in specific domains.
• Brave Search: Find web links for content scraping.
• Fetch: Use fetch_txt and fetch_markdown to obtain web content and convert it into structured data.
• OpenRouter: Analyze topic sentiment and comments.

📚 Documentation

Configuration Instructions

Add the following content to the MCP client configuration file:

"roo-code-memory-bank-mcp": {
  "autoapprove_tools": ["initialize_memory_bank", "check_memory_bank_status", "read_memory_bank_file", "append_memory_bank_entry"],
  "command": "node index.js"
}

Project Context Maintenance

Always use the roo-code-memory-bank-mcp server to save the project status:

1. At the Start of a Task
   • Call check_memory_bank_status to confirm the existence of the context.
2. Load Context Information
   • Use read_memory_bank_file to obtain the content of relevant files, ensuring that the AI assistant understands the current project background.
3. Record Key Decisions and Progress
   • Update the decisionLog.md or progress.md file through append_memory_bank_entry to maintain project memory.
4. Initialize a New Project
   • If the context does not exist and it is a new project, use initialize_memory_bank to create the initial structure.