MemOS provides a robust, private, and local-first memory layer for agents to remember every interaction without cloud dependency.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memos-oneclick-install
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install memos-oneclick-install using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is MemOS: Persistent Local Memory?
MemOS is a high-performance local memory plugin designed specifically for the ecosystem of Openclaw Skills. It ensures that every conversation is automatically captured, semantically indexed, and instantly recallable, all while keeping data 100% on-device. By utilizing a local SQLite database, it eliminates the need for cloud uploads or external API keys for core functionality, making it a cornerstone for privacy-conscious developers and power users.
Beyond basic recall, MemOS transforms agent interactions into structured tasks and evolves successful executions into reusable skills. This allow agents to improve over time, sharing knowledge across multi-agent teams via its Hub-Client architecture. With its built-in Memory Viewer, users can visually inspect and manage the agent's long-term knowledge base, making it one of the most powerful Openclaw Skills for creating truly intelligent assistants.
MemOS: Persistent Local Memory Use Cases
- Maintaining long-term conversation history and context across multiple sessions.
- Automatically summarizing complex coding or research tasks for future reference.
- Distilling successful agent workflows into reusable skills that auto-upgrade over time.
- Setting up collaborative, shared memory for multi-agent teams using Hub-Client architecture.
- Developing privacy-first AI applications where data must remain strictly on the local machine.
How MemOS: Persistent Local Memory Works
- Conversation turns are automatically captured via the agent_end lifecycle hook.
- Content is semantically chunked and indexed into a local SQLite database located at ~/.openclaw/memos-local/memos.db.
- Relevant past memories are searched and injected into the agent's context before every turn using the before_agent_start hook.
- Completed tasks are analyzed and summarized into structured entries for high-level retrieval.
- Successful executions are evolved into skills and stored in the local skill store for future use.
- Users can manage, browse, and edit all indexed data through the local Memory Viewer web UI.
MemOS: Persistent Local Memory Setup
To install this as part of your Openclaw Skills library, follow these steps:
# 1. Install the plugin (standard method)
node -e "process.env.MEMOS_SKIP_SETUP='1';require('child_process').execSync('openclaw plugins install @memtensor/memos-local-openclaw-plugin',{stdio:'inherit'})"
# 2. Rebuild native modules for your specific Node.js version
node -e "require('child_process').execSync('npm rebuild better-sqlite3',{cwd:require('path').join(require('os').homedir(),'.openclaw', 'extensions', 'memos-local-openclaw-plugin'),stdio:'inherit'})"
# 3. Update configuration to disable built-in search and enable the plugin
# (The agent usually handles the JSON patching for you)
# 4. Restart the gateway to apply changes
openclaw gateway restart
MemOS: Persistent Local Memory Data Schema & Taxonomy
| Component | Description | Default Path |
|---|---|---|
| Database | Persistent SQLite store for all indexed memories | ~/.openclaw/memos-local/memos.db |
| Viewer Auth | Configuration and authentication for the web UI | ~/.openclaw/memos-local/viewer-auth.json |
| Plugin Code | Extension directory containing core logic | ~/.openclaw/extensions/memos-local-openclaw-plugin/ |
| Skills Store | Directory for auto-generated and evolved skills | ~/.openclaw/memos-local/skills-store/ |
| Memory Guide | Skill documentation for memory operations | ~/.openclaw/workspace/skills/memos-memory-guide/ |
MemOS: Persistent Local Memory Advanced Features
- Hub-Client Multi-Agent Memory: Share memories and skills across different agents and local network clients.
- Skill Evolution Engine: Automatically identifies successful task patterns and distills them into reusable agent skills.
- Custom Embedding Providers: Support for external APIs including OpenAI, Gemini, Cohere, Voyage, and Mistral for improved search quality.
- LLM-Powered Task Summarization: Automatically structures long conversations into concise, actionable summaries.
- Visual Memory Viewer: A full-featured web interface at http://127.0.0.1:18799 for manual data management and inspection.
SKILL.md
Loading
Related Openclaw Skills
Backlink Prospector Lite
An AI-driven assistant that automates guest post opportunity discovery and generates personalized outreach emails to scale your link-building efforts.
MCP Security Auditor Lite
A specialized security tool designed to identify critical vulnerabilities within Model Context Protocol server configurations.
Client Proposal Generator Lite
A streamlined business development tool designed to transform project briefs into professional executive summaries and high-level scope outlines.
TRPG Session Framework
A multi-agent framework designed to run immersive tabletop RPG campaigns with AI-driven Dungeon Masters and Player Characters.
AIButton
A specialized generator for creating Vision Node AI button controls with integrated metadata and interactive logic.
Alibaba Cloud Content Moderation (Green)
A professional toolset for managing Alibaba Cloud Content Moderation resources via OpenAPI and SDK integrations.
Featured*
