Memory Bank

This Memory Bank MCP Server helps you build a structured documentation system based on a Memory Bank pattern for preserving context in AI assistant environments. It provides ready-made templates and analysis to help you organize core documents and contextual files used by AI agents like Cursor.

How to use

You can run Memory Bank MCP Server locally or remotely and connect it to your MCP client. Start by choosing a running method that fits your workflow: UVX, Smithery, Docker, or a manual Python setup. Once running, you can request the server to describe the Memory Bank file structure, generate templates for new Memory Bank files, and analyze your project to suggest Memory Bank content. Use the server to bootstrap consistent documentation for your AI context needs and to streamline ongoing content updates.

Typical usage patterns include: forming a foundation with a core file like projectbrief.md, generating additional context files, and iteratively updating your active and system context as your project evolves. You can ask for a detailed structure description, request a memory bank template for a specific file, or analyze a project summary to receive tailored Memory Bank content suggestions.

How to install

Prerequisites: you need Python installed for the manual setup, and you should be able to run Node-based tools or Docker if you choose UVX, Smithery, or Docker methods.

{
  "mcpServers": {
    "mcp_memory_bank": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/ipospelov/mcp-memory-bank",
        "mcp_memory_bank"
      ]
    }
  }
}

Option 2: Smithery (local CLI) configuration shown here uses npx to run the package with your Smithery key.

{
  "mcpServers": {
    "memory_bank": {
      "command": "npx",
      "args": [
        "-y",
        "@smithery/cli@latest",
        "run",
        "@ipospelov/mcp-memory-bank",
        "--key",
        "your_smithery_key"
      ]
    }
  }
}

Option 3: Docker run configuration.

{
  "mcpServers": {
    "memory_bank": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "19283744/mcp-memory-bank:latest"
      ]
    }
  }
}

Option 4: Manual setup using Python. First create and activate a virtual environment, then install dependencies.

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

{
  "mcpServers": {
    "memory_bank": {
      "command": "python",
      "args": ["src/mcp_memory_bank/main.py"]
    }
  }
}

Additional sections

Available Memory Bank actions include getting the file structure, generating templates for new Memory Bank files, and analyzing a project to suggest Memory Bank content. These can be requested through your MCP client after the server starts.

Available tools

get_memory_bank_structure

Returns a detailed description of the Memory Bank file structure.

generate_memory_bank_template

Returns a template for a specific Memory Bank file.

analyze_project_summary

Analyzes a project summary and provides suggestions for Memory Bank content.