Sno Memory for OpenClaw

Start a new OpenClaw conversation and most agents have forgotten the last one. Sno Memory keeps it: an encrypted memory on your own machine that your OpenClaw agents write to and read from across conversations.

What you get

• Agents that remember. Relevant memories are recalled automatically, and the agent can store, look up and open entries itself with memory_store, memory_recall and memory_read.
• Mistakes get corrected, not erased. memory_correct writes the right text as a new entry and marks the old one retired. Explicit recall shows retired history with its replacement; automatic context only shows what is current. The model has no delete action.
• One memory for every agent. Codex, Claude Code, Hermes Agent and OpenClaw read the same ~/.sno/settings.json, so what one learns is there for the others.
• Yours, and it stays yours. The store is encrypted from first use with a key made once on your machine. Sno never receives your database or your key. There is no daemon to babysit: the plugin starts the memory service when it needs it.
• No account, no API key. The default mode runs on your machine. After a one-time model download during setup, recall works offline. It works in English, Chinese, Japanese, Korean, German, French, Spanish and Russian.

Get started

1. Follow the memory service setup. It takes a few minutes and runs once for every agent on the machine. You need Node.js 22.22.3+, 24.15.0+ or 25.9.0+.

2. Install the plugin:

   openclaw plugins install @snoai/[email protected]

   To upgrade later, run sno update; OpenClaw refuses a second openclaw plugins install while its gateway is running.

3. Restart OpenClaw, tell an agent "Remember that I prefer tabs for indentation", start a new conversation and ask what you prefer.

Walkthrough: OpenClaw onboarding. Daily use and troubleshooting: usage guide.

Modes

Local First (what the memory service setup selects) keeps memory work on your machine with no model call. Agent Native uses the OpenClaw agent's own model, and REM Enhanced adds Sno's own memory models to REM, the overnight step where your agents dream: it strengthens their memory and settles what they learned. It runs at most once a day for each part of your memory: usually overnight, or earlier the same day once about 100 new memories have piled up. Sno's models also write down what matters from each conversation when the session ends. The agent's model does the rest. In that mode your memory text and lessons go to your own space in Sno's cloud, which works like your own cloud drive: a Sno program processes it for you alone, and you can delete or keep it as you like. The recall and capture groups in ~/.sno/settings.json set recall limits and what gets captured. Keep the existing encryption key when you edit it.

Under the hood

This package is the OpenClaw side only: it registers the memory tools and the sno-mem command. The memory itself lives in @snoai/memory, a local service on 127.0.0.1 that owns the encrypted SQLite store, vector and keyword search, and the local embedding model. Uninstalling the plugin does not erase the store.

License

Licensed under Apache-2.0.