OptMem user reports a small edge over Claude Code's native memory
After five weeks with OptMem, a user says memory helped mainly in long sessions by avoiding rediscoveries and time-wasting mistakes, with only small savings on token costs.
TLDR
A user compared five weeks with OptMem against their previous three months using Claude Code's native memory and reported a small improvement. They also saw sessions where Claude didn't activate OptMem, giving them what they described as a kind of A/B test. Their takeaway: memory helps a bit, mostly in long sessions, especially by avoiding rediscoveries and time-wasting mistakes. They intend to keep using OptMem and praise its fixed-size memory for being structured and easy to read.
Combined views
—
2 Sources, first seen ago
