Recherche Memoire

Searches the agent's persistent memory by semantic similarity. Allows retrieving information stored in previous sessions.

Parameters

Parameter	Type	Required	Variable	Description
`query`	dynamic value	Yes	Yes	Text or question to search in the agent long-term memory.
`limit`	number	No	No	Maximum number of results to return. (Default: `10`)
`minScore`	number	No	No	Minimum similarity score (between 0 and 1). Only results above this threshold are returned. (Default: `0.7`, min 0, max 1)
`entityId`	dynamic value	No	Yes	Filter results by entity identifier.
`tags`	array	No	No	Filter results by tags.
`dateFrom`	dynamic value	No	Yes	Start date for the time range filter (ISO 8601).
`dateTo`	dynamic value	No	Yes	End date for the time range filter (ISO 8601).
`includeMetadata`	boolean	No	No	Include metadata in the results. (Default: `false`)
`sortBy`	choice (`score`, `date`, `importance`)	No	No	Result sorting criterion: by similarity score, by date, or by importance. (Default: `"score"`)
`outputVariable`	text	No	No	Output variable name containing the retrieved memories.

Parameters marked Variable = Yes accept the {{blockName.field}} syntax.

Output variable : memorySearchResult

{
  "memories": [],
  "totalCount": 0,
  "query": "...",
  "hasVectorSearch": false
}

Search the agent's memory.

Input :

{"query": "preferences client Acme"}

Output :

{"results": [{"key": "client-acme", "content": "Prefere les livraisons le mardi", "score": 0.88}], "count": 1}

Tip

Useful in agent workflows to recall context from previous sessions. Requires that data has been saved via memory_save.