MemOS Cloud Skill

English | 中文

MemOS Cloud Server API skill. This skill allows Agents or developers to directly call the MemOS Cloud Platform API to retrieve, add, delete, and feedback on memories.

Prerequisites

• Python: 3.x and above
• Python Dependencies: requests module (pip3 install requests)

Install

Option A — Command Line (Recommended)

npx skills add https://github.com/MemTensor/MemOS-Cloud-Skill

Option B — Manual Install

1. Clone this repository to your local machine:

   git clone https://github.com/MemTensor/MemOS-Cloud-Skill.git
   

2. Manually copy the skill folder to your corresponding agent skills directory.

Environment Variables

This step is critical. You must configure these variables before using the skill.

Where to configure

• You can configure these globally in your system environment (e.g., ~/.bashrc, ~/.zshrc).
• Or, you can configure them within your specific AI Agent or framework's environment settings (e.g., .env files for OpenClaw/Moltbot/Clawdbot).

Required

• MEMOS_API_KEY (required; Token auth) — get it at MemOS API Console
• MEMOS_USER_ID (required; A deterministic user-defined personal identifier, e.g., hashed email, employee ID) — Do not use random or session IDs.

MEMOS_API_KEY=YOUR_TOKEN
MEMOS_USER_ID=YOUR_USER_ID

Optional config

• MEMOS_CLOUD_URL (default: https://memos.memtensor.cn/api/openmem/v1)

Quick setup (shell)

echo 'export MEMOS_API_KEY="mpg-..."' >> ~/.bashrc
echo 'export MEMOS_USER_ID="user-123"' >> ~/.bashrc
source ~/.bashrc

Quick setup (Windows PowerShell)

[System.Environment]::SetEnvironmentVariable("MEMOS_API_KEY", "mpg-...", "User")
[System.Environment]::SetEnvironmentVariable("MEMOS_USER_ID", "user-123", "User")

How it Works / Usage

Once installed and configured, this skill empowers your AI Agent (e.g., Trae, Cursor, OpenClaw) to manage your long-term memories autonomously. Simply communicate with your Agent through natural language, and it will intelligently decide when to call the underlying MemOS APIs based on your conversations.

1. Add Message (/v1/add/message)

When you share preferences, facts, or instructions you want the Agent to remember, it will automatically extract the high-value content and save it to the MemOS cloud.

Example Conversation:

• You: "Please remember that my primary programming language is Python and I prefer dark mode."
• Agent: (Recognizes intent -> Calls add_message skill) "Got it! I've saved your preferences about Python and dark mode."

2. Search Memory (/v1/search/memory)

Before answering complex questions or when explicitly asked, the Agent will search your past memories to provide highly personalized responses.

Example Conversation:

• You: "Write a boilerplate script for my usual tech stack."
• Agent: (Recognizes intent -> Calls search skill to retrieve your python preferences) "Sure! Here is a set of Python boilerplate code..."

3. Delete Memory (/v1/delete/memory)

If a memory is outdated or incorrect, simply tell the Agent to forget it.

Example Conversation:

• You: "Forget my previous residential address, I've moved."
• Agent: (Recognizes intent -> Calls delete skill) "I have removed your old address from my memory."

4. Add Feedback (/v1/add/feedback)

You can correct the Agent's behavior, and it will reinforce its memory for future interactions.

Example Conversation:

• You: "Your last answer wasn't detailed enough. Next time, always provide code comments."
• Agent: (Recognizes intent -> Calls add_feedback skill) "Understood. I will add more details and code comments in the future."