Llama.cpp, opencode / pi / basically all agents, context compaction & cache validation: how do you manage it?

r/LocalLLaMA
Generative AI Open Source AI

Ok so, I will try to explain myself as much as possible because onlinew I really cannot find much about this.