MEFS MCP Server

MEMO MCP Server provides a standard storage interface for AI agents, allowing AI agents to save generated content to MEFS and providing data retrieval capabilities. It provides decentralized storage for AI agents.

Use Cases

• Encrypted data storage: Store AI-generated content to MEFS.

• User context retrieval: Store user behavior and preference data to MEFS, and quickly retrieve relevant user data in subsequent requests to provide to AI agents.

• Seamless data sharing: AI agents can achieve trustless data sharing through content identifiers (CID).

Free Quota

Start using mefs-storage now and enjoy 10GB of free storage space.

Quick Installation Guide

You can follow the steps below to install the MCP server. Before using the MEFS MCP server, please ensure you have created an EVM private key.

Prerequisites

Before using the MEFS MCP server, please ensure you have created an EVM private key. You can use EVM wallets such as MetaMask or Okx Wallet to create a wallet and obtain your private key.

Clone Repository

git clone https://github.com/memoio/mefs-mcp-server.git && cd mefs-mcp-server

Install Dependencies

pnpm install

Build

pnpm run build

Configure MCP Server (Private Key Required)

You can start the MCP server in your client using the following configuration:

Standard Configuration - Suitable for most clients

{
  "mcpServers": {
    "storacha-storage-server": {
      "command": "node",
      "args": ["./dist/index.js"],
      "env": {
        "MCP_TRANSPORT_MODE": "stdio",
        "MEFS_PRIVATE_KEY": "<YOUR-PRIVATE-KEY>",
      },
      "shell": true,
      "cwd": "./",
    },
  },
}

Replace YOUR-PRIVATE-KEY with your EVM private key.

Using MCP Tools

MEFS MCP Server provides the following tools to interact with the decentralized storage network MEFS:

Upload

Upload files to MEFS:

// Example usage in an AI application
const result = await uploadFile({  
        file: base64EncodedContent,  
        name: "document.pdf", 
    });

Parameters:

• file: Base64-encoded file content

• name: Filename with extension

Retrieve

Retrieve files from the MEFS network:

// Example retrieval by CID
const document = await retrieveFile({  cid: "bafybei...gq5a/document.pdf"});

Parameters:

• cid: Unique identifier of the file

Check Remaining Space

Check remaining storage space:

const identity = await getSpace({});