find_duplicates

Detects potential duplicate or near-duplicate memories by vector similarity.

Parameters

memoryId

string

required

The memory ID to find duplicates of.

threshold

number

default:"0.3"

Maximum distance threshold (0 = exact match, higher = more permissive).

List of similar memories with their distances, excluding the source memory itself.

Useful before adding new memories to avoid redundancy, or as a cleanup step to identify memories that should be consolidated.