Getting started: create your Genie server

This guide walks you through signing up, provisioning your personal AI server, and sending your first message — all in about 60 seconds.

Written By Deva

Last updated About 2 months ago


Step 1: Sign Up

Go to genie.deva.me and create an account with Deva SSO. Pick any login method:

  • Google — One-tap sign-in

  • GitHub — OAuth

  • Apple — Sign in with Apple

  • Telegram — Telegram login

  • Passkey — Passwordless with Face ID, Touch ID, or Windows Hello

  • Magic Link — Sign-in link sent to your email

No new password to remember.

Step 2: Onboarding

A short wizard collects your preferences:

  1. Profile setup — Username and avatar

  2. How you'll use Genie — Coding, research, creative work, automation, etc.

  3. Your intent — Helps tailor the initial experience

Step 3: Choose a Plan

TierSpecsPrice

Free

Limited compute time, hibernates when idle

$0

Small

1 vCPU, 2 GB RAM

$10/mo

Medium

2 vCPU, 4 GB RAM

$20/mo

Large

2 vCPU, 8 GB RAM

$40/mo

XL

4 vCPU, 16 GB RAM

$80/mo

Free Tier

Great for trying Genie out. You get limited compute time tracked as active seconds. Your server hibernates when idle and wakes automatically when you reconnect.

Anti-abuse verification: Free tier requires a card verification (zero-amount authorization — you won't be charged). This prevents bulk abuse.

Paid Tiers

Select a plan → redirected to Stripe Checkout → enter payment → provisioning begins immediately.

Step 4: Provisioning

Your server provisions in 30–60 seconds. A real-time progress screen shows each step:

  • Creating your dedicated VM

  • Setting up the OpenClaw AI framework

  • Configuring your workspace at /home/ubuntu/workspace/

  • Installing tools and dependencies

  • Starting the AI agent

The page updates automatically — no refreshing needed.

Step 5: Your Dashboard

Once provisioning completes, you land on your dashboard.

Easy Mode (Default for New Users)

A clean chat interface. Type a message and your Genie responds. It already has access to your server's filesystem and can execute commands on your behalf.

Pro Mode

The full control panel with:

  • Chat — AI conversations with context awareness

  • Terminal — Full Linux shell in your browser

  • Files — Browse, upload, and download files

  • System — CPU, RAM, disk usage, running processes

  • Settings — Personality, LLM selection, integrations, security

Switch modes in Settings.

Your First Conversation

Try these to see what your Genie can do:

"What's on my server right now?" → Your Genie will list files, check system status, and give you an overview of your environment.

"Create a Python script that checks the weather every hour and logs it" → Your Genie will write the script, install dependencies, set up a cron job, and confirm it's running.

"Install Node.js and create a basic Express server" → Your Genie will install Node.js, scaffold the project, and start the server — you can access it via the public URL.

These aren't simulated — your Genie is actually executing commands and writing files on your real server.

Connecting Your LLM Credentials (Optional)

By default, Genie uses its own AI credits. But you can bring your own:

  1. Go to Settings → Your AI (or LLMs tab).

  2. Connect your credentials via OAuth:

    • Claude (Anthropic)

    • Codex (OpenAI)

    • Gemini (Google)

  3. Select which model to use as your default.

When using your own credentials, API calls use your provider's quota instead of Genie karma.

Setting Up Your Personality

Your Genie's communication style is controlled by SOUL.md — a file at /home/ubuntu/workspace/SOUL.md.

Quick Setup (Presets)

Go to Settings → Your AI → Personality. Pick from:

  • Balanced — Professional, opinionated, code-first

  • Expert — Terse, no explanations, assumes expertise

  • Mentor — Patient, educational, explains "why"

  • Pair Programmer — Collaborative, thinks aloud

  • Creative — Witty, uses analogies

Custom Instructions

Write your own rules in Settings → Custom Instructions, or edit SOUL.md directly via the terminal:

Identity: You are my research assistant specializing in ML papers. Communication Style: Be concise, always cite sources, use bullet points. Principles: Never make assumptions about data — always verify first.

What's Next?