Onboarding Wizard

Interactive setup guide for first-time configuration

What the Wizard Does

The onboarding wizard guides you through:

1. AI Provider Selection - Choose between local (Ollama) or cloud providers
2. Model Setup - Download and configure your AI model
3. Gateway Configuration - Set port, authentication, and network settings
4. Channel Setup - Connect messaging apps (optional)
5. Workspace Configuration - Set your working directory

Running the Wizard

Step-by-Step Guide

Step 1: AI Provider Selection

Choose your AI provider:

• Ollama (Local) - Recommended for privacy, runs offline, zero cost
• OpenAI - GPT-4, GPT-3.5-turbo
• Anthropic - Claude models
• OpenRouter - Access multiple models
• Custom - Any OpenAI-compatible API

Step 2: Model Setup

For Ollama:

• Wizard checks if Ollama is installed
• Installs Ollama if missing
• Shows available models to download
• Downloads your selected model (e.g., llama3.2)

For Cloud Providers:

• Enter your API key
• Select default model
• Test connection

Step 3: Gateway Configuration

Configure the gateway service:

• Port - Default: 18789
• Authentication - Token or password
• Network - Localhost, LAN, or Tailscale

Step 4: Channel Setup (Optional)

Connect messaging channels:

• Telegram - Enter bot token
• Discord - Enter bot token
• WhatsApp - QR code pairing
• Skip to add channels later

Step 5: Workspace

Set your workspace directory (default: ~/.jarvis/workspace)

Re-running Onboarding

You can re-run onboarding anytime to:

• Add new AI providers
• Change configuration
• Add more channels
• Reset settings