Tool reference · Knowledge graph & temporal facts
memory_concepts
Lists or searches nodes in the knowledge graph by name, optionally with the memories linked to each one.
local stdio server read-only
When to use
- You want to know which concepts the memory already knows about.
- You need the exact node name before calling memory_graph.
- You want to see which saved records are attached to a concept.
Parameters
| Name | Type | Default | Description |
|---|---|---|---|
query | string | — | Search concepts by name |
type | string | — | Filter by node type |
limit | integer | 20 | — |
include_memories | boolean | false | Include linked knowledge records |
Example
Arguments
{
"query": "auth",
"limit": 5,
"include_memories": false
}Result shape
{
"concepts": [
{
"id": "n_4a2e",
"type": "concept",
"name": "authentication",
"importance": 0.64,
"mention_count": 38
},
{
"id": "n_4a31",
"type": "concept",
"name": "oauth",
"importance": 0.41,
"mention_count": 12
}
],
"total": 2
}Without query it lists nodes of the given type, defaulting to "concept". With include_memories each node gets a memories array.
Values are illustrative; the keys follow the server's handler. MCP clients receive the result as JSON text content.
Server description
The description the server sends to your agent in tools/list, captured from the v14.7.0 source:
List or search concepts in the knowledge graph.