Tool reference · Knowledge graph & temporal facts
memory_associate
Finds memories by spreading activation through the knowledge graph from the concepts in your query, rather than by keyword match.
local stdio server writes
Caution. Concepts are matched only against names already in the graph. If none match, memories comes back empty.
When to use
- Keyword search misses results that are related by topic but worded differently.
- You want to assemble a solution from several past memories that together cover all the concepts you need (mode "composition").
- You want to see which graph concepts your query maps to.
Parameters
| Name | Type | Default | Description |
|---|---|---|---|
queryrequired | string | — | Natural language query |
mode | "recall" | "composition" | "recall" | recall=find related, composition=build solution from parts |
project | string | — | Filter by project |
max_results | integer | 10 | — |
min_coverage | number | 0.7 | Min coverage for composition mode (0.0-1.0) |
Example
Arguments
{
"query": "retry webhooks without duplicate orders",
"mode": "recall",
"project": "my-api",
"max_results": 5
} Result shape
{
"query_concepts": [
"webhooks",
"retry"
],
"seed_nodes": [
{
"id": "n_2b7f",
"name": "webhooks"
},
{
"id": "n_2c01",
"name": "retry"
}
],
"activated_count": 46,
"memories": [
{
"id": 812,
"type": "solution",
"content": "Use an Idempotency-Key header check before creating orders.",
"project": "my-api",
"activation_score": 0.83
}
],
"composition": null
} In composition mode, composition holds sources, coverage_percent, covered, gaps, conflicts and integration_plan.
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:
Associative recall — brain-like spreading activation through knowledge graph. Finds memories through concept resonance, not keyword search. In 'composition' mode, finds minimum set of memories covering all needed concepts.