Memory Bank MCP Server

Remote memory bank management for multi-project file structure and validation.

Data last scanned yesterday · Reviewed today

425 tools★ 917TypeScriptMIT

Overview

Memory Bank MCP Server provides a centralized interface for managing project-specific memory bank files. By implementing the Model Context Protocol, it allows AI assistants to remotely access, read, and update documentation files while enforcing a consistent directory structure across multiple projects. The server is designed to manage file isolation through path traversal prevention and project-specific directory validation. It supports core operations including listing available projects, viewing individual file contents, and updating existing memory bank documentation within defined root directories.

Our verdict

This server receives a security score of 42/100, resulting in a high risk level. The lack of an editorial review and the fact that it has not received a code update in over 351 days contribute to this rating. Given the lack of verified security maintenance, this tool is best suited for experimental environments or non-sensitive projects where local isolation is strictly managed by the user.

Supports multiple project-specific directories for memory bank storage.
Features path traversal prevention to isolate project data.
Operates with only two direct dependencies for lower supply-chain surface.
Maintains 917 GitHub stars within the developer community.

Last code commit occurred 351 days ago.
Not maintained by an official vendor.
Read-only and OAuth implementations are currently unreviewed.

Setup

Claude Desktop

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

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), then restart Claude Desktop. Some servers require API keys or tokens as environment variables — check the project's README and add an "env" object if needed.

No read-only mode — running this grants full read/write access.
No built-in authentication — anyone with access to this config can invoke every tool it exposes.

Tools

Tool	Description	Risk
memory_bank_read	Read memory bank files	low
memory_bank_write	Create new memory bank files	medium
memory_bank_update	Update existing memory bank files	medium
list_projects	List available projects	low
list_project_files	List files within a project	low

Compatibility

Client	Local	Docker	Remote	Read-only
Claude Desktop
Cursor
VS Code
Windsurf
ChatGPT

Frequently asked questions

›What is the purpose of the MEMORY_BANK_ROOT environment variable?

The MEMORY_BANK_ROOT variable defines the local directory where your project memory bank files are stored.

›Can I use this server to manage multiple projects?

Yes, the server supports multi-project management by enabling file structure enforcement and proper isolation between different project memory banks.

›Does the server allow me to automate file operations without repeated prompts?

Yes, you can configure an autoApprove list in your MCP settings to bypass manual approval for specific operations like reading, writing, and updating memory bank files.

›Is it possible to run this MCP server inside a Docker container?

Yes, the server supports Docker deployment by mapping your local memory bank directory to the container and configuring the command arguments accordingly.

›Where should I configure the MCP server for the Cline or Roo Code extensions?

You need to add the configuration to the specific settings JSON file located in your user's globalStorage directory for those extensions.