Remote MCP Server

Remote MCP Server

Enables AI tools like Claude to interact with a remote machine's file system and shell via a secure HTTPS endpoint. It provides standardized tools for executing shell commands, reading and writing files, and navigating directories.

Category
访问服务器

README

mcp-remote-server

A lightweight MCP (Model Context Protocol) server that runs on a remote machine, giving AI tools like Claude access to that machine over HTTPS.

What it does

Exposes four tools via a single /mcp HTTP endpoint:

Tool Description
run_command Run a shell command on the remote server
read_file Read a file from the remote server
write_file Write content to a file on the remote server
list_directory List contents of a directory on the remote server

Quick start

The automated installer handles everything — user creation, SSL certificates, systemd service, and firewall detection:

git clone https://github.com/OddbeakerLLC/mcp-remote-server.git
cd mcp-remote-server
sudo bash setup.sh

The installer will:

  1. Create a dedicated mcp-server system user
  2. Detect whether port 443 is available
    • Free → obtains a Let's Encrypt certificate and runs HTTPS directly
    • In use → runs HTTP on port 3098 and prints a reverse proxy config snippet for your web server
  3. Create and enable a systemd service with auto-restart
  4. Print your connector URL

Connecting to claude.ai

This server is designed to be used as an MCP connector on claude.ai:

  1. Open claude.ai and go to Settings
  2. Navigate to Connectors (or MCP Connectors)
  3. Click Add connector
  4. Enter your server URL: https://your-server.example.com/mcp
  5. Save — Claude will now have access to your remote machine's tools

Connecting to Claude Code

Add this to your Claude Code MCP settings (~/.claude.json or project-level .mcp.json):

{
  "mcpServers": {
    "my-remote-server": {
      "type": "url",
      "url": "https://your-server.example.com/mcp"
    }
  }
}

Manual setup

If you prefer not to use the installer, or want to customize the setup:

Standalone with Let's Encrypt

# Get a certificate
sudo certbot certonly --standalone -d your-server.example.com

# Run the server
SSL_CERT=/etc/letsencrypt/live/your-server.example.com/fullchain.pem \
SSL_KEY=/etc/letsencrypt/live/your-server.example.com/privkey.pem \
PORT=443 \
node server.js

Behind nginx

Run the server on an internal port:

PORT=3098 node server.js

Add to your nginx config:

server {
    listen 443 ssl;
    server_name your-server.example.com;

    ssl_certificate     /path/to/fullchain.pem;
    ssl_certificate_key /path/to/privkey.pem;

    location / {
        proxy_pass http://127.0.0.1:3098;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
    }
}

Behind Caddy

Caddy handles SSL automatically. Run the server on port 3098, then add to your Caddyfile:

your-server.example.com {
    reverse_proxy localhost:3098
}

Environment variables

Variable Description Default
PORT Port to listen on 443 (with SSL) / 3098 (without)
SSL_CERT Path to SSL certificate (fullchain.pem) (none — runs HTTP)
SSL_KEY Path to SSL private key (privkey.pem) (none — runs HTTP)

See .env.example for a template.

Security warnings

This server gives an AI agent shell access to your machine. Take precautions:

  • Create a dedicated user account with limited permissions. Do not run this as root.
  • Firewall the port so only trusted IPs can reach it.
  • Always use HTTPS. Never expose this over plain HTTP on the internet. The installer handles this automatically.
  • Understand the risk: any prompt injection or model mistake can run arbitrary commands on your server.
  • Network isolation: consider running this on an internal network or behind a VPN for additional security.

Health check

curl https://your-server.example.com/health

Returns JSON with server status and uptime.

Systemd service management

If you used the installer, the service is managed via systemd:

# Check status
systemctl status mcp-remote-server

# View logs (live)
journalctl -u mcp-remote-server -f

# Restart
systemctl restart mcp-remote-server

# Stop
systemctl stop mcp-remote-server

Troubleshooting

Server won't start — "SSL certificate not found" Check that SSL_CERT and SSL_KEY point to valid files. If using Let's Encrypt, verify the cert exists at /etc/letsencrypt/live/yourdomain/.

Port 443 permission denied Binding to ports below 1024 requires root or the CAP_NET_BIND_SERVICE capability. The systemd service file sets this automatically. For manual runs, use sudo or run on port 3098 behind a reverse proxy.

certbot fails to obtain a certificate Make sure your domain's DNS A record points to your server's public IP, and that port 80 is open (certbot needs it for the HTTP challenge).

Claude can't connect

  • Verify the server is running: curl https://your-domain.com/health
  • Check firewall rules allow inbound traffic on port 443
  • Ensure your SSL certificate is valid (not expired, matches the domain)

License

MIT

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选