Navy AI 🚢

Navy AI is a terminal-based AI assistant designed for local-first usage, with optional cloud providers.
It works entirely from the command line and prioritizes privacy, cost control, and clarity.

🟢 Default mode is FREE and OFFLINE using local AI.

---

✨ Features

• 🖥️ Clean, modern CLI
• 🧠 Local AI via Ollama (free, offline)
• ☁️ Cloud AI via Gemini (free tier available)
• 💳 Optional OpenAI support (paid, opt-in)
• 🔁 Argument mode & interactive mode
• 🎨 Styled terminal output
• 🔌 Extensible provider system
• 🔐 Secure by default (no keys in code)

---

📦 Installation

Requirements

• Python 3.9+
• (Optional) Ollama for local AI

Install from PyPI

pip install navy-ai

---

📘 Usage Guide

Verify Installation After installing Navy AI, verify that the CLI is available:

navy-ai --help

---

🚀 Quick Start

Navy AI works in two modes:

• Argument mode – single command
• Interactive mode – chat-style session

🔹 Argument Mode

Ask a question directly from the terminal:

navy-ai "what is zero trust security?"

Example output:

Zero Trust is a security model that assumes no implicit trust...

---

🔹 Interactive Mode

Start an interactive session:

navy-ai

You will see:

Navy AI >

Then type your questions:

Navy AI > what is a cpu
Navy AI > explain zero trust
Navy AI > exit

---

🧠 Providers Overview

Provider	Cost	Internet	Notes
Ollama	Free	❌ No	Local, offline, recommended
Gemini	Free tier	✅ Yes	Google AI Studio
OpenAI	Paid	✅ Yes	Requires billing

🟢 Default provider is Ollama (local-first).

Ollama (Local AI – Recommended)

Ollama allows you to run AI models locally and offline.

1️⃣ Install Ollama

👉 https://ollama.com

2️⃣ Pull a Model

ollama pull mistral
ollama pull qwen2.5-coder:7b
ollama pull llama3
..........

3️⃣ Use with Navy AI

Explicit provider + model:

navy-ai --provider ollama --model mistral "what is cpu"

Or simply:

navy-ai "what is cpu"

➡️ Ollama is the default provider.

🟡 Gemini (Cloud AI – Free Tier)

1️⃣ Create an API Key

Keys must be created from Google AI Studio:

👉 https://aistudio.google.com/app/apikey

⚠️ API keys from Google Cloud Console will not work.

2️⃣ Set Environment Variable

Windows (PowerShell)

setx GEMINI_API_KEY "AIzaSyXXXX"

macOS (Terminal) Set the variable (temporary – current session only):

export GEMINI_API_KEY="AIzaSyXXXX"

Make it persistent (recommended): For zsh (default on modern macOS):

echo 'export GEMINI_API_KEY="AIzaSyXXXX"' >> ~/.zshrc

For bash:

echo 'export GEMINI_API_KEY="AIzaSyXXXX"' >> ~/.bashrc

Restart the terminal (or run source ~/.zshrc / source ~/.bashrc).

Verify:

echo $GEMINI_API_KEY

Linux (Terminal)

Set the variable (temporary – current session only):

export GEMINI_API_KEY="AIzaSyXXXX"

Make it persistent:

For bash:

echo 'export GEMINI_API_KEY="AIzaSyXXXX"' >> ~/.bashrc

For zsh:

echo 'export GEMINI_API_KEY="AIzaSyXXXX"' >> ~/.zshrc

Restart the terminal (or run source ~/.bashrc / source ~/.zshrc).

Verify:

echo $GEMINI_API_KEY

3️⃣ Use Gemini

navy-ai --provider gemini

Recommended model:

navy-ai --provider gemini --model gemini-2.5-flash

🔵 OpenAI (Optional – Paid)

OpenAI requires billing to be enabled.

1️⃣ Create API Key

👉 https://platform.openai.com/api-keys

2️⃣ Enable Billing

👉 https://platform.openai.com/account/billing

2️⃣ Set Environment Variable

Windows (PowerShell)

setx OPENAI_API_KEY "sk-xxxx"

macOS (Terminal) Set the variable (temporary – current session only):

export OPENAI_API_KEY="sk-xxxx"

Make it persistent (recommended): For zsh (default on modern macOS):

echo 'export OPENAI_API_KEY="sk-xxxx"' >> ~/.zshrc

For bash:

echo 'export OPENAI_API_KEY="sk-xxxx"' >> ~/.bashrc

Restart the terminal (or run source ~/.zshrc / source ~/.bashrc).

Verify:

echo $OPENAI_API_KEY

Linux (Terminal)

Set the variable (temporary – current session only):

export OPENAI_API_KEY="sk-xxxx"

Make it persistent:

For bash:

echo 'export OPENAI_API_KEY="sk-xxxx"' >> ~/.bashrc

For zsh:

echo 'export OPENAI_API_KEY="sk-xxxx"' >> ~/.zshrc

Restart the terminal (or run source ~/.bashrc / source ~/.zshrc).

Verify:

echo $OPENAI_API_KEY

4️⃣ Use OpenAI

navy-ai --provider openai --model gpt-3.5-turbo "explain zero trust"

⚠️ If billing is not enabled, OpenAI may return:

429 Too Many Requests

⚙️ CLI Syntax

navy-ai [OPTIONS] [PROMPT]

Options

Option	Description
--provider	ollama
--model	Provider-specific model
--help	Show help

🧪 Examples

navy-ai "hi!"

navy-ai --provider ollama --model qwen2.5-coder:7b

navy-ai --provider gemini --model gemini-2.5-flash

navy-ai --provider openai --model gpt-3.5-turbo