Knowledge Graph Memory Server

You manage a persistent knowledge graph memory server that stores entities, their observations, and the relationships between them. This MCP server lets you create, link, tag, search, and retrieve nodes and observations, enabling structured memory for projects, teams, and workflows. It runs as an MCP endpoint you can query from your client, responding with structured data about your knowledge graph.

How to use

Connect your MCP client to the knowledge graph server to manage entities, relations, and observations. You can create multiple entities, attach observations, define relations between entities, and tag entities or observations for easy discovery. Use search to locate nodes by name, type, or content of observations, then open and inspect targeted nodes and their connections.

How to install

Prerequisites: you need Node.js and npm installed on your machine. You will run the knowledge graph MCP server as an MCP endpoint using npx.

# Install prerequisites if you don't have Node.js and npm
# On macOS with Homebrew
brew install node

# On Ubuntu/Debian
sudo apt-get update
sudo apt-get install -y nodejs npm

Additional sections

Configuration and runtime notes: the server runs via a local MCP command using npx and the knowledge graph MCP package. You specify the path to your persistent database file when starting the server. The database file will be created automatically if it does not exist, so choose a stable location to store your graph data.

Usage patterns: create entities with types, attach observations, and add relations to describe how entities interact. Tag entities and observations to enable fast discovery and filtering. Use read, search, and open operations to explore the graph and understand dependencies and context.

Available tools

create_entities

Create multiple new entities in the knowledge graph. Input is an array of entities with name, entityType, and observations.

create_relations

Create multiple new relations between entities. Input is an array of relations with from, to, and relationType.

add_observations

Add new observations to existing entities. Input includes entityName and contents (array of observations) and returns the added observations per entity.

delete_entities

Remove entities and their associated relations with cascading deletion. Input is an array of entityNames.

delete_observations

Remove specific observations from entities. Input includes entityName and observations to delete.

delete_relations

Remove specific relations from the graph. Input includes from, to, and relationType.

read_graph

Read and return the complete knowledge graph, including all entities and relations, with no input.

search_nodes

Search for nodes based on a text query across entity names, types, and observation content. Returns matching entities and their relations.

open_nodes

Retrieve specific nodes by name and return the entities plus the relations between them. Non-existent nodes are silently skipped.

tag_entity

Attach tags to an entity. Creates tags if they do not exist and returns the list of successfully added tags.

tag_observation

Attach tags to a specific observation. Creates tags if they do not exist and returns the list of successfully added tags.

get_entities_by_tag

Find entities that have a specified tag and include their related relations.

get_all_tags

List all available tags, including their categories and descriptions.

get_tag_usage

Provide usage statistics for tags across entities and observations.

remove_tags_from_entity

Remove specific tags from an entity and return the list of successfully removed tags.