Alice-Ai

A local terminal-based AI assistant powered by Ollama…

A local terminal-based AI assistant powered by Ollama with lightweight persistent memory stored in SQLite.
Designed for fast CLI interaction, experimentation, and local automation workflows.

This project runs entirely locally. No cloud APIs, no external telemetry, no credentials.

---

Attribution & Disclosure

This project was partially written with the assistance of the Ollama 2.3.8B model.

I do not claim full authorship of all logic, structure, or ideas contained in this repository.
The code represents a collaborative output between human direction, iteration, and AI-assisted generation.

All responsibility for usage, modification, and deployment remains with the user.

---

Features

• Terminal chat interface using a local Ollama model
• Optional memory injection for contextual continuity
• Structured memory system with categories
• SQLite-backed local storage
• Streaming responses
• Simple command-based control

---

How Memory Works

Memory is stored locally in a SQLite database:

~/.alice/alice_memory.db

Only predefined categories are allowed to keep memory structured:

• identity
• preferences
• goals
• workflow
• skills
• projects
• constraints

If an invalid category is provided, it defaults to preferences.

Memory can be:

• Pinned (always injected first)
• Normal (recently injected until limit)

A size cap is enforced to prevent prompt bloat.

---

Requirements

• Linux or macOS (recommended)
• Python 3.10+
• Ollama installed and running locally
• At least one Ollama model pulled

Python dependencies:

• ollama
• rich

---

Installation

Clone the repository:

git clone

https://github.com/siv-the-programmer/Alice-Ai.git  
cd Alice-Ai

Create and activate a virtual environment:

python3 -m venv .venv  
source .venv/bin/activate

Install dependencies:

pip install -U pip  
pip install ollama rich

---

Running the Bot

Start the CLI:

python3 CLI_bot.py

On first run, the application will take 1 to 2 mins to fully start as it will automatically create:

~/.alice/
~/.alice/alice_memory.db

No manual setup required.

Screenshot1

Screenshot2

---

CLI Commands

- /help — Show commands  
- /mem — List memories  
- /save <category>|<text> — Save memory  
- /pin <category>|<text> — Save pinned memory  
- /forget <id> — Delete memory  
- /reset — Reset database  
- /chaos <1–10> — Adjust response intensity  
- exit — Quit  

Examples:

/save projects|Building Python automation tools  
/pin workflow|Be concise and ask one clarification if unsure  
/chaos 8

---

GitHub Safety

This repository contains no secrets or credentials.

Excluded by design:

• SQLite databases
• User memory
• Environment files

All personal data remains local and is ignored by Git.

---