All ClawAll Claw
Back to Catalog
MemOS Cloud OpenClaw Plugin logo
skills

MemOS Cloud OpenClaw Plugin

Official lifecycle plugin for OpenClaw that integrates MemOS Cloud memory, enabling persistent long-term recall and conversation saving to reduce token usage and improve agent context across runs.

Overview

The MemOS Cloud OpenClaw Plugin is the official lifecycle plugin developed by MemTensor for integrating MemOS Cloud memory capabilities into OpenClaw agents. It allows agents to recall relevant memories from the MemOS Cloud before each execution and automatically add new messages and context to the memory layer after completion.

This plugin addresses common pain points in agentic systems like high token consumption from repeated context and lack of persistent memory across sessions, making it ideal for long-running, multi-task agents.

Features

  • Automatic Recall: Hooks into the before_agent_start lifecycle to fetch and inject relevant memories via MemOS Cloud /search/memory endpoint.
  • Conversation Persistence: On agent_end, adds new messages and outcomes to MemOS Cloud using /add/message.
  • Token Efficiency: Significantly reduces prompt token usage by avoiding redundant history in every run (reported reductions up to 70% in benchmarks).
  • Simple Authentication: Uses token-based auth with your MemOS API key.
  • Minimal Overhead: Lightweight JavaScript implementation designed specifically as an OpenClaw lifecycle plugin.
  • Cross-Agent Sharing: Enables multiple OpenClaw instances to share a common memory layer for collaborative or multi-agent workflows.

Installation and Setup

  1. Install the plugin via npm or directly from GitHub.
  2. Configure your MemOS Cloud API key in the plugin settings.
  3. Add it to your OpenClaw agent's plugin configuration (memory slot).
  4. The plugin activates automatically on agent lifecycle events.

Refer to the README for detailed setup instructions and environment variables.

Use Cases

  • Long-Running Agents: Maintain context over days or weeks without bloating prompts.
  • Multi-Agent Systems: Allow agents to share and build upon collective knowledge.
  • Token-Cost Optimization: Ideal for production deployments where LLM API costs are a concern.
  • Persistent Skill Reuse: Store and recall learned skills or patterns across different tasks.
  • Personal AI Assistants: Keep conversation history and user preferences persistent.

Community and Support

Maintained by MemTensor, the team behind MemOS. Join discussions on GitHub issues, or explore the broader MemOS project for advanced memory features. Active development with recent commits and growing adoption in the OpenClaw community.

Tags

openclawpluginmemoryai-agentlong-term-memorymemosmemos-cloudjavascript