Getting Started

Go from zero to a working AI assistant in 5 minutes with minimal setup.

Prerequisites

  • macOS, Linux, or Windows (via WSL2)
  • 8GB+ RAM (16GB recommended for local AI)
  • 10GB free disk space

Quick Setup (CLI)

The fastest way to get started is with our one-line installer:

curl -fsSL https://nildesai.com/install.sh | bash

This installer will:

  1. ✅ Install Node.js 22+, Git, and pnpm (if missing)
  2. 🤖 Install Ollama for local AI
  3. 📥 Download a default AI model (llama3.2 - 4GB)
  4. 🔧 Clone and build Jarvis from source
  5. 🚀 Launch the interactive onboarding wizard

What You'll Have

  • A running Gateway process
  • Local AI with Ollama (no API keys needed)
  • Web dashboard at http://127.0.0.1:18789
  • CLI tools ready to use

First Commands

After installation, try these commands:

Test your local AI:

jarvis agent --message "Hello, Jarvis!" --local

Start the gateway:

jarvis gateway

Open the web dashboard:

open http://127.0.0.1:18789

Check system health:

jarvis doctor

Useful Environment Variables

Customize Jarvis behavior with these environment variables:

VariableDefaultDescription
JARVIS_CONFIG_DIR~/.jarvisConfig directory
JARVIS_GIT_DIR~/jarvisSource code location
OLLAMA_HOST127.0.0.1:11434Ollama server address

Next Steps

Jarvis

© 2026 Nilkanth Desai. MIT licensed.