zettel-compress

Deterministic memory engine for LLM apps — compress, search, and inject conversation memory offline, on-device, or at the edge. Zero external calls, zero infrastructure, zero dependencies.