Last released Jul 2, 2026
Smart context compression for LLM agents — preserve critical info, reduce tokens 40-80%
Supported by