Dedicated memory sub-agent before the main reply; modes, overrides, and optional transcript debugging.
Active Memory (optional plugin)
Active Memory runs a dedicated memory sub-agent immediately before the main reply so ongoing chats can pull in preferences, context, and past details without you repeating “remember this” every time.
Modes & tuning
- Configurable message, recent, and full context modes
/verbose-style inspection for debugging- Optional transcript persistence for troubleshooting
Docs: upstream Active Memory concept page (linked from the official OpenClaw docs).