zettel-compress playground

deterministic LLM memory — compress, search, inject. Everything runs in this tab. No server, no model, no deps.

conversation / document

ask the memory

zettels

Paste text on the left (or load the sample) — compression runs live as you type.

budget 300 tokens