with
Enable eager topic-based similarity search.
Uses the topic field to perform a vector similarity search and preloads the top matching memories into the description. This makes the most relevant memories for the topic immediately visible to the LLM without a tool call.
Subsequent tool calls automatically deduplicate against these eagerly loaded memories, so the LLM always receives new information.
Can be combined with withEagerQuery — both sets of memories will be merged (deduplicated) in the description.
Return
New Memory with eager topic search configured
Parameters
limit
Maximum number of memories to preload (default DEFAULT_LIMIT)