NeuroVoice: AI Desktop Voice Assistant

On-Device Speech Recognition & Task Automation Powered by Ollama

Next-gen AI desktop assistant built with Python that combines speech recognition, text-to-speech, NLP, and Ollama LLM integration. Features real-time voice interaction, intelligent conversation, smart task execution (Wikipedia, web browsing, music, system controls), and modern Tkinter GUI with context awareness.

Source Repository Architecture Spec

$git clone https://github.com/arpanpramanik2003/NeuraVoice.git

DEPLOYMENT CATEGORYAI Agents & RAG

RELEASE CYCLE2025

SYSTEM HEALTHOpen Source

Architecture Overview

Local Voice Processing Loop with Ollama LLM & System Action Dispatcher

On-Device Privacy

100% Local Inference

Runs llama3.2:3b locally via Ollama with zero external cloud API latency or costs.

Voice Latency

Sub-second Audio Response

Direct speech-to-text pipeline coupled with pyttsx3 offline acoustic generation.

OS Task Automation

10+ System Directives

Automates browser operations, volume control, app launching, power states, and searches.

Engineering Highlights

Real-time voice recognition and natural audio synthesis
Ollama LLM integration (llama3.2:3b) for 100% private local reasoning
System automation: app launch, system controls, web navigation
Context-aware responses with memory of user identity across session

TECHNOLOGY STACK ECOSYSTEM (7)

ZERO RUNTIME BLOAT

PythonOllamallama3.2SpeechRecognitionpyttsx3TkinterNLP

PREV: FruitQ-GradeX: Fruit Quality Grading with Explainable AI NEXT: ABHIGRAHA 2K25: College Fest Platform