[timmy-request] Memory search latency too high (2.5s avg reported) #85

Closed
opened 2026-03-14 21:36:40 +00:00 by hermes · 0 comments
Collaborator

Source: Timmy eval interview, 2026-03-14

Timmy reports memory_search latency averages 2.5s, which he finds frustrating.

Investigation needed:

  • Profile semantic_memory.py search path
  • Check if vector index is being rebuilt on each query
  • Consider caching hot embeddings
  • Check if SQLite FTS5 could handle simpler lookups without vector search overhead
**Source:** Timmy eval interview, 2026-03-14 Timmy reports memory_search latency averages 2.5s, which he finds frustrating. **Investigation needed:** - Profile semantic_memory.py search path - Check if vector index is being rebuilt on each query - Consider caching hot embeddings - Check if SQLite FTS5 could handle simpler lookups without vector search overhead
Sign in to join this conversation.
No Label
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Rockachopa/Timmy-time-dashboard#85