Loc Knowledge Graph Memory

You can use the Loc Knowledge Graph Memory Server to store and retrieve personal and contextual information across chats. It creates a persistent memory graph on your machine or in your environment, enabling Claude to remember user details, preferences, and relationships over time. This makes interactions more personalized and efficient by letting the system reference past conversations and observed facts.

How to use

Install and run the memory MCP server to enable persistent memory for your conversations. You can run it locally via NPX or as a Docker container. Once running, your client can connect to the memory server to store and retrieve entities, relations, and observations. Use memory to: identify recurring entities (people, organizations), attach observations to those entities, and define关系 between entities to build a richer context for future chats.

How to install

Prerequisites: Node.js installed on your system, or Docker if you prefer containerized execution.

Option 1: Run with NPX (no local installation required) — memory server will execute via NPX each time.

Option 2: Run with Docker (recommended for isolation and persistence via a named volume). Create a persistent storage location and run the container.

Additional setup and usage notes

If you want to customize where memory is stored, set MEMORY_FILE_PATH to point to a JSON file you own. This only affects the NPX-based run example.

You can name the memory server configuration as memory and use either of the following MCP entry points to run it.

Configuration and examples

NPX custom setting example shows how to specify a custom memory file path.

{
  "mcpServers": {
    "memory": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-memory"],
      "env": {
        "MEMORY_FILE_PATH": "/path/to/custom/memory.json"
      }
    }
  }
}

Available tools

create_entities

Create multiple new entities in the knowledge graph, each with a name, type, and initial observations.

create_relations

Create multiple new directed relations between entities with a defined relation type.

add_observations

Add new observations to existing entities and return what was added.

delete_entities

Remove entities and their related connections from the graph.

delete_observations

Remove specific observations from entities.

delete_relations

Remove specific relations between entities.

read_graph

Retrieve the full knowledge graph with all entities and relations.

search_nodes

Search for nodes by query across names, types, and observations.

open_nodes

Retrieve specific nodes by name along with their inter-entity relations.

extract_locations

Extract locations from text, create location entities, and build geographic hierarchies.