All servers / Memory (knowledge graph)
Memory (knowledge graph)
Persistent knowledge-graph memory for agents.
Tested by ConnectATDScore 100 out of 100. Tested 2026-11-17 on package version 2.1.1. Package: @modelcontextprotocol/server-memory · source
Score breakdown
| Runs | 20 / 20 | |
|---|---|---|
| Tool descriptions | 20 / 20 | |
| Package safety | 20 / 20 | |
| Maintenance | 20 / 20 | |
| Adoption | 20 / 20 |
What we saw
- Starts and answers the MCP handshake: yes (3.0 seconds including download)
- Tools offered: 9
- Last published: 2026-11-11
- Weekly downloads (npm): 384,991
- License: SEE LICENSE IN LICENSE · Direct dependencies: 2
Notes
- Safe test call to
read_graph: worked.
Tools and how they are described
Agents choose tools from these descriptions. Under each name you can see how many of its parameters are explained.
create_entities1 of 1 parameter describedCreate multiple new entities in the knowledge graph. An entity whose name already exists, or repeats an earlier entity in the same call, is skipped and its observations are not added; the result lists the skipped names in `skipped`. Use add_observations to add observations to an
create_relations1 of 1 parameter describedCreate multiple new relations between entities in the knowledge graph. Relations should be in active voice
add_observations1 of 1 parameter describedAdd new observations to existing entities in the knowledge graph
delete_entities1 of 1 parameter describedDelete whole entities from the knowledge graph, together with every relation that points to or from them and all their observations. Use it only to remove an entity entirely; to remove a link between two entities and keep both, use delete_relations, and to remove facts from an en
delete_observations1 of 1 parameter describedDelete specific observations from entities in the knowledge graph
delete_relations1 of 1 parameter describedDelete specific relations (from, to, relationType) from the knowledge graph and keep the entities they connect. Each relation must match an existing one exactly on all three fields; if unsure, call open_nodes (or search_nodes) first to read the stored relationType. Use it to remo
read_graph2 of 2 parameters describedRead the entire knowledge graph: every entity with its observations, and every relation. The result grows with the graph, so use this only when you need everything; prefer search_nodes to find entities by keyword and open_nodes to read entities whose names you know.
search_nodes3 of 3 parameters describedSearch the knowledge graph for entities whose name, type or observations contain the query (case-insensitive). Use this when you have a keyword, a topic or part of a name but not the exact entity names. Prefer open_nodes when you already know the exact names, and use read_graph o
open_nodes2 of 2 parameters describedRetrieve specific entities by their exact names, with every relation that touches one of them. Use this when you already know the names (from the conversation, or from an earlier search_nodes or read_graph result); it is the cheapest way to read entities you can name. Prefer sear
Tested in a sandbox with an empty environment: no API keys, a throwaway folder, nothing from a real account. A passing test shows the server runs and responds. It is not a security audit.