A pure-Python context management layer for LLM systems — retrieval, re-ranking, memory decay, and token-budget enforcement in one pipeline.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).