xgmem MCP Memory Server

Provides a persistent, cross-project memory store with entities, relations, and observations for MCP clients.
  • typescript

5

GitHub Stars

typescript

Language

2 months ago

First Indexed

3 weeks 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": {
    "meetdhanani17-xgmem": {
      "command": "npx",
      "args": [
        "-y",
        "xgmem@latest"
      ],
      "env": {
        "MEMORY_DIR_PATH": "/app/memories"
      }
    }
  }
}

xgmem is a TypeScript-based MCP server that provides project-specific memory and knowledge-graph storage for entities, relations, and observations. It helps agents and tools store and retrieve structured memory across projects, enabling cross-project sharing and scalable, disk-persistent memory for your agent ecosystem.

How to use

You interact with xgmem through MCP clients that send tool calls to the server. Use the provided tools to save observations, create entities and relations, read and search knowledge graphs, and copy or migrate memory between projects. Start the server locally, then connect your MCP client to the stdio runtime to perform memory operations such as saving project observations, querying project memories, and updating the memory graph.

How to install

Prerequisites you need before installing: Node.js and npm are installed on your system. Ensure you have a compatible npm and Node.js version for your environment.

#!/usr/bin/env bash
# Install dependencies for the memory server project
npm install

# Build the project (if a build step is exposed by the project)
npm run build

Configuration and usage notes

Configure the MCP server entry in your MCP configuration to run xgmem as a local stdio server. This uses npx to run the latest xgmem package, so you can keep the server up-to-date with minimal changes.

{
  "mcpServers": {
    "xgmem": {
      "command": "npx",
      "args": ["-y", "xgmem@latest"]
    }
  }
}

Run (development)

Run the memory server in development mode using the provided runtime. This keeps you in a rapid feedback loop while you develop or test memory operations.

npx ts-node index.ts

Run (production)

When you are ready for production, start the server with the production command to ensure a stable runtime.

npm start

Docker

Use Docker to run xgmem with persistent memory storage on the host. The memory files are stored under the memories directory on the host.

docker build -t xgmem-mcp-server .
docker run -v $(pwd)/memories:/app/memories xgmem-mcp-server

Available tools

save_project_observations

Saves memory observations for a specific project, storing structured observations under the given projectId.

get_project_observations

Retrieves observations for a specific project, enabling review of stored memories.

add_graph_observations

Adds new observations to the existing memory graph, enhancing entity graphs and relations.

create_entities

Creates new memory entities within a project context.

create_relations

Defines new relations between entities to extend the knowledge graph.

delete_entities

Removes specified memory entities from a project.

delete_observations

Deletes observations associated with a project or specific entities.

delete_relations

Removes relations between entities in a project memory.

read_graph

Reads and returns the current memory graph for a project, including entities and their relations.

search_nodes

Searches nodes within the memory graph by name or attributes.

search_all_projects

Searches memory across all projects to identify shared or related nodes.

open_nodes

Opens and fetches detailed information for specific memory nodes.

copy_memory

Copies memory data between projects or within a project for migration or backup.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational