Memory Bank

Memory Bank MCP Server provides persistent project knowledge across sessions using MCP to store global and branch-specific memory.

typescript

11

GitHub Stars

typescript

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": {
    "t3ta-memory-bank-mcp-server": {
      "command": "yarn",
      "args": [
        "workspace",
        "@memory-bank/mcp",
        "start",
        "--docs",
        "/path/to/your/docs"
      ]
    }
  }
}

Memory Bank MCP Server provides a centralized, structured memory store for project documentation and context that lasts across sessions. It helps AI agents keep consistent knowledge through global and branch-specific memory banks and exposes operations via MCP so you can manage, query, and evolve memory over time.

How to use

You run the Memory Bank MCP Server from a local development workspace and connect your MCP client to it. The server stores global memory and branch-specific memory in a structured JSON format so your agents can retrieve relevant context for long-running conversations, summaries, and decision logs. Use the MCP client to perform memory creation, retrieval, updates, and memory-tree navigation, enabling your AI to stay aligned with project knowledge across sessions.

How to install

# Prerequisites
#  - Node.js (refer to your toolchain versioning, e.g., .tool-versions)
#  - Yarn v1.x

# Install dependencies from the root of the workspace
git clone https://github.com/t3ta/memory-bank-mcp-server.git
cd memory-bank-mcp-server
yarn install

# Start the MCP server from the monorepo
# Replace /path/to/your/docs with the actual docs directory for your project
yarn workspace @memory-bank/mcp start --docs /path/to/your/docs

Additional notes

This server is part of a monorepo managed with Yarn Workspaces. It includes the core MCP implementation and related schemas for memory documents. If you are developing locally, you can build and test packages with the available scripts in the workspace and start the MCP server using the command shown above.