Getting Started with NubiferAI

Getting Started

Installation

Standalone Installation

# Install from PyPI
pip install nubiferai

# Or install in a virtual environment
python3 -m venv ~/.local/nubiferai-venv
source ~/.local/nubiferai-venv/bin/activate
pip install nubiferai

NubiferOS Installation

If you're running NubiferOS, install NubiferAI as an add-on:

nubi addon install nubiferai

This automatically configures workspace integration, credentials, and safety classification.

System Requirements

• Python 3.11 or later
• GTK4 and libadwaita (for the GUI application)
• An AI provider (see Providers)

GTK4 Dependencies (Linux)

# Ubuntu/Debian
sudo apt install python3-gi gir1.2-gtk-4.0 gir1.2-adw-1 libadwaita-1-dev

# Fedora
sudo dnf install python3-gobject gtk4 libadwaita-devel

# Arch
sudo pacman -S python-gobject gtk4 libadwaita

First Launch

GUI Application

nubiferai-gtk

On first launch you'll see:

1. Header bar — Workspace indicator, safety classification, provider/model selectors
2. Status bar — Shows whether you're in standalone mode or connected to NubiferOS
3. Prompt area — Where you describe what you want to do
4. Quick Actions — Pre-configured buttons grouped by category (AWS, GCP, Azure, etc.)

If no API key is configured, you'll see a "Standalone Mode" banner with a link to Settings.

CLI

# Check your environment
nubiferai status

# List available seeds
nubiferai seed list

# Generate your first nucleation
nubiferai nucleate "Create an S3 bucket with versioning enabled"

Configuring a Provider

Before generating nucleations, you need at least one AI provider configured. The quickest options:

Free: Ollama (Local)

Run AI models on your own machine — no API key, no cost:

# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh

# Pull a model
ollama pull llama3.2

# Set as your provider
nubiferai config --init
# Select "ollama" when prompted

Paid: Anthropic Claude (Recommended)

Best quality for infrastructure planning:

# Set your API key
export ANTHROPIC_API_KEY="sk-ant-your-key-here"

# Add to your shell profile for persistence
echo 'export ANTHROPIC_API_KEY="sk-ant-your-key-here"' >> ~/.bashrc

See Providers for all options including OpenAI and AWS Bedrock.

Your First Nucleation

Via GUI

1. Select a provider and model from the header bar dropdowns
2. Optionally click a Quick Action (e.g., "Deploy to AWS") to pre-select a seed
3. Type your request: "Create a VPC with public and private subnets in us-east-1"
4. Press Ctrl+Enter or click Generate
5. Review the generated nucleation — actions grouped by operation type
6. Click View Code to see the Terraform/shell output
7. Click Export to save to a file
8. Click Precipitate to execute (with safety confirmation)

Via CLI

# Generate a nucleation
nubiferai nucleate "Create a VPC with public and private subnets in us-east-1"

# Use a specific seed
nubiferai nucleate --seed terraform-expert "Refactor my ECS cluster to use Fargate"

# Export to Terraform
nubiferai nucleate --seed aws-deployer "Deploy Redis on ElastiCache" --export terraform

# Execute directly (with confirmation)
nubiferai front

Configuration

NubiferAI configuration is stored at ~/.config/nubiferai/config.toml:

provider = "claude"
default_seed = "aws-deployer"
require_confirmation = true

[safety]
classification = "development"
can_plan = true
can_apply = true
require_confirmation = true

Configuration Hierarchy

Settings are resolved in this order (later overrides earlier):

1. System defaults — built-in defaults
2. User config — ~/.config/nubiferai/config.toml
3. Workspace config — .nubiferai/config.toml in workspace root
4. Environment variables — NUBIFERAI_PROVIDER, etc.
5. NubiferOS overrides — workspace safety classification (if on NubiferOS)
6. Command-line flags — --provider, --seed, etc.

Safety Classification

NubiferAI enforces safety rules based on the environment classification:

On NubiferOS, the safety classification is set automatically based on your workspace configuration.

Next Steps

• Set up your preferred provider
• Explore the GUI features
• Browse available seeds
• Learn about NubiferOS integration