Tool reference · Maintenance & performance
memory_warmup
Loads the embedding model and opens the vector store now, so the first real save or search is not slow.
local stdio server writes idempotent
When to use
- Right after the memory server starts, before the first user request.
- The first search of a session is noticeably slower than the rest.
- You want to check which vector backend is active.
Parameters
This tool takes no parameters.
Example
Arguments
{}Result shape
{
"fastembed_loaded": true,
"vector_backend": "sqlite_binary",
"ms": 1840
}vector_backend is sqlite_binary, chroma or none.
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:
pre-load FastEmbed model and open the vector store, so the first save/search after process start doesn't pay model-load latency.