Condense.chat

A compression proxy for AI coding-agent workflows that sits between an agent and its model, removing unnecessary tokens and compacting settled agent loops to reduce context and token use.