Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/Seaus-tech/Aurora-Shell/llms.txt

Use this file to discover all available pages before exploring further.

Auseaia (pronounced aw-SAY-ah) is Aurora Shell’s built-in AI assistant. It runs entirely on your machine using the Llama3 model through Ollama, which means no API keys, no cloud round-trips, and no data leaving your system. You can open an interactive multi-turn chat session or fire off a one-shot question directly from your terminal. Conversation history is persisted between sessions and can be cleared at any time with /reset.

​
Prerequisites

Before using Auseaia, Ollama must be installed and the Llama3 model must be available locally.
1

Install Ollama

Run the automated setup script, or install Ollama manually:
# Automated (recommended)
bash ~/.aurora-shell_files/Aurora-Shell/setup_ai.sh

# Manual install
curl -fsSL https://ollama.com/install.sh | sh
2

Start the Ollama service

The setup script starts Ollama in the background automatically. If you installed manually, start it yourself:
ollama serve
3

Pull the Llama3 model

The setup script pulls the model automatically. To do it manually:
ollama pull llama3

​
Using the AI Assistant

​
Interactive chat session

Launch a persistent multi-turn conversation: 
auseaia

# Alternatively, run the chat script directly
bash ~/aurora-shell-2/chat.sh
Type your message at the ❯ prompt and press Enter. Use /reset to clear the conversation history, or type exit (or quit) to end the session.

​
One-shot query

Pass your question as a command-line argument for a single response without entering the interactive interface: 
auseaia "explain what a shell alias is"

auseaia "write a bash function to find large files"

​
The Chat Interface

When you open an interactive session, Auseaia displays a header panel followed by the command reference and the input prompt: 
╭─────────────────────────────────────────╮
│  🌌 Auseaia - Local AI Assistant      │
│  Powered by Llama3 via Ollama         │
╰─────────────────────────────────────────╯
Commands: /reset (clear history) | exit (quit)

❯

​
Conversation History

All messages — including system context, your inputs, and Auseaia’s replies — are written to history.json in the aurora-shell repo directory after each exchange. When you start a new session, Auseaia loads this file so context carries over from previous conversations. To wipe the history and start fresh, type /reset at the prompt. This overwrites history.json with an empty array and prints a confirmation message.

​
System Prompt

Auseaia is initialized with the following system-level instruction at the start of every fresh conversation:
“You are Auseaia, a helpful AI assistant running locally via Llama3. You help with coding, analysis, and terminal tasks. Be concise and direct.”
This prompt is injected automatically the first time a message is sent in a clean session — you do not need to configure it manually.

​
Streaming Responses

Auseaia uses the Ollama streaming API so the model’s reply appears token-by-token as it is generated, rather than waiting for the full response before printing anything. While the model is processing, a braille spinner animates in place: 
⠋ Thinking...
The spinner frames cycle through ⠋ ⠙ ⠹ ⠸ ⠼ ⠴ ⠦ ⠧ ⠇ ⠏ at 80 ms intervals. As soon as the first token arrives, the spinner is cleared and the response begins streaming to the terminal.
Auseaia runs 100% locally — no data is sent to any external service. Requires Ollama to be running in the background (ollama serve). If Ollama is not running when you invoke auseaia, you will see: ❌ Error: Unable to connect to Ollama. Make sure it's running with 'ollama serve'

Build docs developers (and LLMs) love

Get started for freeTalk to us