Memory plugin for OpenClaw that runs 100% locally. Smarter search, memory quality checks, and compaction rescue — so your agent remembers everything, no matter how long the conversation. No API keys. No subscriptions. Just install and go.
$49 — one-time purchase, no subscriptionThree features that make long conversations actually work.
Search works instantly with BM25 text matching — no embedding provider configuration required. OpenClaw's built-in memory needs you to set up an embedding provider (OpenAI, Gemini, etc.) for vector search. Memory Stack gives you fast, accurate search out of the box, plus optional local vector search via QMD with zero external API calls.
Built-in health checks that find duplicate memories, stale entries with expired dates, and noise. Other plugins just pile up data forever. Memory Stack scores your memory quality and tells you what to clean up — it never deletes anything automatically.
Long conversations get compressed by OpenClaw — and key decisions get lost in the process. Memory Stack automatically saves important facts (decisions, deadlines, requirements) before they're compressed away, and brings them back when you need them.
Memory Stack plugs into OpenClaw as a native memory provider. No manual setup after install.