A
license-
qualityD
maintenanceEnables 70-90% LLM API cost reduction by compressing conversation history via local Gemma 4 models or heuristics, featuring token counting, model routing, and pinned facts for preserving critical context.
Last updated
1
MIT