# OpenClaw Setup + MEMO MEFS Agent Service: Complete Usage Guide

*Mastering the OpenClaw Setup: A Comprehensive Guide to Leveraging MEMO MEFS Services*

By [MemoLabs](https://paragraph.com/@memolabs) · 2026-03-31

---

**I. Prerequisites**

Setting Up Your Environment

To run the MEFS MCP Server on Windows, you’ll need to install two foundational tools: **Node.js** (the JavaScript runtime) and **Git** (for downloading the source code).

**Install Node.js:**

Visit [Node.js — Node.js 24.14.1 (LTS)](https://nodejs.org/en/blog/release/v24.14.1?ref=blog.memolabs.org).

Select the appropriate version, download it, and run the installer. Keep all default options and click “Next” through each screen to complete the installation.

![](https://storage.googleapis.com/papyrus_images/133c1e886fbca6995955e243ea49da30c06e2436f8e0a7c440dbd63ffc360022.png)

**Install Git:**

Visit [Git — Install for Windows](https://git-scm.com/install/windows?ref=blog.memolabs.org).

Download the Windows version and install it.

![](https://storage.googleapis.com/papyrus_images/0f0b12a8ae0a03424338f3b02a12d3faae41756ef7c07cd5d1572c14df57a82a.png)

Verify the Installation:

Press **Win + R**, type **cmd**, and press Enter to open a Command Prompt.

Run **node -v** and **git — version**. If both commands return version numbers, the installations were successful.

![](https://storage.googleapis.com/papyrus_images/d1cd0d3b5d527a344171e2469b349a9a0fd248255f15682b048115e5931e752f.png)

**II. Installing and Setting Up the OpenClaw Runtime**

1\. Click the Windows Start menu, search for Windows PowerShell, and open it.

2\. Download and install the OpenClaw runtime

Run this command: **iwr -useb** [**https://openclaw.ai/install.ps1**](https://openclaw.ai/install.ps1) **| iex**

Note: Installation takes approximately 10 minutes. The terminal will display “**Installing OpenClaw**” throughout the process — please be patient and do not close the window.

![](https://storage.googleapis.com/papyrus_images/44027849f7797f7171c54f13d013f406f7f6409bb799bc00e0594ecebcb8e90e.png)

3.When you see the prompt “I understand this is personal-by-default and shared/multi-user use requires lock-down. Continue?”:

l Use the **Up/Down arrow keys (↑ / ↓)**.

l Move the cursor to **Yes** (the confirmation option, typically listed above “No”).

l Press **Enter**.

![](https://storage.googleapis.com/papyrus_images/487e9b03e7bf3673294c1939fe2c659c8a853027e17faec5b16e887225fbc3a5.png)

4\. When the startup mode selection screen appears, choose “Quick Start”.

![](https://storage.googleapis.com/papyrus_images/622f2125e0dac4d7a9e82c02619e1996bb2c828da8f4a2b78e9145841c03a9fb.png)

5\. Select the AI model you want to use from the list.

![](https://storage.googleapis.com/papyrus_images/879f7adf11fd0f3db35313f833a98550ced52eb5b6906e114ad8d3557f37e037.png)

6\. Enter the API key for your selected model.

![](https://storage.googleapis.com/papyrus_images/51a44e365f498521e4607cf7d58bbaada149f92cc3e0173159abe6203edf3946.png)

You can skip all remaining options.

![](https://storage.googleapis.com/papyrus_images/9fcf0fedea92d450be797134b0e39a6accb4ffb5b339dec00280854426cff214.png)

Configure Skills: We recommend selecting “No” here — it’s much easier to install Skills from the web interface later.

![](https://storage.googleapis.com/papyrus_images/67a6077a521e926a590672fd1f02bb573266b20fe6aec3c29326e16fb9ac6877.png)

Note: The only required configuration is the API key

**III. Install MEFS MCP Server**

Open a new Windows PowerShell window.

**1.** **Clone the repository**

Run: git clone [https://github.com/memoio/mefs-mcp-server.git](https://github.com/memoio/mefs-mcp-server.git)

**2.** **Navigate into the directory**

Run: **cd mefs-mcp-server**

**3.** **Install dependencies**

Run: **pnpm install**

![](https://storage.googleapis.com/papyrus_images/68eae9d795209fe61e3535517b878d472ed71fdf04b4bde00c1d8391b102ae88.png)

4\. Build the project

Run: **pnpm build**

![](https://storage.googleapis.com/papyrus_images/37261bbae5cf6e4cb5c450ace49d2e9c866ce2486a6e18e1c1f18a11280b39dc.png)

**IV. Start the server**

Have your **EVM private key** (without the 0x prefix) ready, then run the following in the current window:

$env:MEFS\_PRIVATE\_KEY=”your\_private\_key”; $env:PORT=8085; $env:MCP\_TRANSPORT\_MODE=”sse”; node dist/index.js

**Success indicator**：: you should see MCP SSE Server running on [http://0.0.0.0:8085/sse](http://0.0.0.0:8085/sse)[.](http://0.0.0.0:8085/sse.?ref=blog.memolabs.org)

![](https://storage.googleapis.com/papyrus_images/ee7dd0bc90e7488ec2a76f686eaab5d8969bd83d7b5c639f823b028cb01e69f4.png)

**V. Putting It to Work**

Once the service is running, open your browser and navigate to the token-authenticated URL shown in the terminal. You can then issue instructions directly to the AI Agent in the chat window.

![](https://storage.googleapis.com/papyrus_images/4b215578891cefb346e45d9002c66a9253b2a6073bf0ffc164acbc19f7a8a43d.png)

![](https://storage.googleapis.com/papyrus_images/7d47e93803f12bdf5ddfde9e155b86596998ad86fb08a181010543c09a20953c.png)

---

*Originally published on [MemoLabs](https://paragraph.com/@memolabs/openclaw-setup-memo-mefs-agent-service-complete-usage-guide)*
