Documentation

The interfaces between you and the cloud.

M.A.F Cloud is a lightweight panel built on top of Docker containers, supporting Minecraft servers, Node.js, Python, Golang, and generic web application hosting. These pages explain what each layer does, where the boundaries are, and how to use them safely.

00

Quickstart

M.A.F Cloud runs the panel; you bring the VPS. Three steps from signup to first server or application deployment.

  1. Sign up at https://mafcloud.my.id/login and confirm your email.
  2. In the dashboard, click Add Node. You'll get a one-line install command tied to your account, e.g.
    Terminal
    curl -fsSL https://cexi.my.id/agent-install.sh \
  | sudo MAFCLOUD_TOKEN=mn_... MAFCLOUD_API=https://cexi.my.id bash
    Run it on your VPS as root (or inside a GitHub Codespace). The agent registers itself with the panel and starts.
  3. Back in the dashboard, click New server, pick your template (Minecraft, Node.js, Python, Nginx, Golang, etc.), set the RAM, and you're done. The panel tells your node to spin up the corresponding Docker container.

A tiny VPS (1 vCPU / 2 GB RAM) on DigitalOcean, Hetzner, or any KVM host works perfectly for small instances.

02

API surface

The backend exposes REST under /api and one WebSocket gateway at /ws/console. All routes except register, login, and GET /health require a Bearer JWT.

The first user to register is automatically promoted to ADMIN. Subsequent users are USER and only see their own servers.

03

AI Agent

M.A.F Cloud AI is a general-purpose assistant built into the dashboard. It can manage your servers (list, start, stop, check logs, browse files) and answer general questions. Each request draws from your plan's token pool.

  • Global AI — accessible from the main dashboard. Can manage all your servers and nodes.
  • Instance AI — per-server agent inside each server detail page. Scoped to just that server.
  • Token pool — FREE: 1M tokens, PRO: 10M tokens, STUDIO: 50M tokens per period.
04

API v1 & MCP Protocol (PRO/STUDIO)

Private OpenAI-compatible API and Model Context Protocol (MCP) server for PRO and STUDIO plans. Authenticate with your API key (mf_live_xxx) or Bearer JWT token.

Terminal
# List models (OpenAI Compatible)
curl -H "Authorization: Bearer mf_live_xxx" https://cexi.my.id/api/v1/models

# Chat completion (OpenAI Compatible)
curl -X POST https://cexi.my.id/api/v1/chat \
  -H "Authorization: Bearer mf_live_xxx" \
  -H "Content-Type: application/json" \
  -d '{"model":"hy3-free","messages":[{"role":"user","content":"Hello"}]}'

# Model Context Protocol (MCP) - List Tools
curl -H "Authorization: Bearer your_jwt_token" https://cexi.my.id/mcp/tools

# MCP - Execute Tool
curl -X POST https://cexi.my.id/mcp/tools/call \
  -H "Authorization: Bearer your_jwt_token" \
  -H "Content-Type: application/json" \
  -d '{"name":"list_servers","arguments":{}}'

Models available: hy3-free, mimo-v2.5-free. MCP tools support full node and server control (list_nodes, control_server, get_latest_logs, list_files, trigger_self_heal).

05

Token System

Every AI request deducts tokens from your plan's allocation. Token usage is tracked per user and resets each plan period.

  • FREE — 300M tokens/period. Dashboard AI only (no API access).
  • PRO — 5B tokens/period. Includes API key access.
  • STUDIO — 100B tokens/period. Unlimited API access.

Check your remaining tokens via GET /api/users/me or the dashboard account page.