Offline AI Voice Assistant Using TinyLLaMA

Overview

This project implements a fully offline AI voice assistant using Python and Ollama. The system listens to user speech, generates intelligent responses using a local language model (TinyLLaMA), and replies using text-to-speech without requiring an internet connection or paid APIs.

---

Workflow

1. Speech Input

User voice input is captured through a microphone and converted into text using the speech_recognition library with optimized microphone sensitivity.

2. Local AI Inference

The recognized text is passed to TinyLLaMA, a lightweight local language model running via Ollama, which generates responses completely offline.

3. Voice Output

The generated response is converted into speech using pyttsx3 and played back to the user in real time.

---

Project Structure

offline-voice-assistant

voice_assistant_offline.py # Main offline voice assistant script README.md # Project documentation requirements.txt # Python dependencies

---

Technologies Used

• Python
• Ollama
• TinyLLaMA
• SpeechRecognition
• Pyttsx3

---

Key Features

• Fully offline AI voice assistant
• No internet connection or API keys required
• Optimized for 4GB RAM systems
• Real-time speech input and voice output
• Privacy-focused local execution

---