Project Deep Dive
Artificial Intelligence Neural-Core v2.60

JARVIS 2.0: Next-Gen
AI Desktop Assistant

A fully functional desktop AI that sees you, hears you, tracks your hand gestures, recognises your face, controls your entire computer, and converses intelligently. It works in real time, built in Python from scratch.

PythonCore Language
GroqAI Backend
v2.60Current Build
10+Core Features
Overview

What is JARVIS 2.0?

JARVIS 2.0 started as a simple voice assistant and grew into something far more ambitious: a desktop AI system that operates across multiple modalities simultaneously. It watches through your camera, listens through your microphone, tracks your hand movements, identifies your face, and responds through a holographic-inspired HUD overlay.

Everything is wired together in Python. The vision pipeline runs MediaPipe for hand landmark tracking and face recognition, feeding into a Groq-powered LLM that processes the context. Edge TTS delivers natural-sounding speech. The system integrates with Spotify, the file system, and browser tabs, meaning it can actually act on your computer, not just talk about it.

"JARVIS isn't just an application. It's the next step in how a human and a computer should interact: one that feels invisible, natural, and genuinely intelligent."

Features

Core Capabilities

Groq Intelligence
Lightning-fast LLaMA 3 inference via Groq API. Responses feel instantaneous, with zero wait time even in complex conversational contexts.
Hand Gesture Control
MediaPipe hand landmark tracking detects custom gestures in real time. Pinch to pause, swipe to switch apps, all without touching anything.
Facial Recognition
Recognises the enrolled user's face on startup. JARVIS greets you by name and adjusts its interaction context based on who is present.
Holographic HUD
A transparent overlay UI inspired by sci-fi interfaces shows system status, AI listening state, and live sensor data in the corner of the screen.
Full System Control
Opens and closes applications, switches browser tabs, manages virtual desktops, adjusts volume, and executes terminal commands, all via voice.
Spotify Integration
Mood-aware playlist control. Tell JARVIS how you feel and it queues a matching playlist, skips tracks, adjusts volume, or searches by artist.
Real-Time Web Search
Tavily API integration allows JARVIS to search the web live and answer questions about current events, news, or anything outside its training window.
Smart Writing
Dictate any text, email, or document verbally. JARVIS opens Notepad (or any target app) and types the content naturally, character by character.
Tech Stack

Built With

Python 3.10.9 Groq API LLaMA 3.3 70B MediaPipe OpenCV Edge TTS face_recognition Tavily Search API Spotify Web API pyautogui SpeechRecognition tkinter (HUD)

Why Python? Python's ecosystem (from MediaPipe to pyautogui to the Groq SDK) meant every component had a library that "just worked." The goal was building something real, not reinventing wheels. Python made that possible for a 14-year-old with no team and no budget.

Evolution

Version History

VersionKey AdditionsStatus
v1.0Basic voice assistant, simple command parsing, pyttsx3 TTSRetired
v1.5Added Groq API backend, faster responses, improved NLPRetired
v2.0Facial recognition, hand gesture control, MediaPipe integrationRetired
v2.30Holographic HUD overlay, Spotify control, system commandsRetired
v2.50Tavily web search, Edge TTS, smart writing modeRetired
v2.60Full feature integration, performance tuning, stable releaseCurrent
v3.0On-device inference, multi-user profiles, gesture vocabulary expansionPlanned

The biggest lesson from building JARVIS was that the gap between "I want to build that" and "I built it" is smaller than most people think. You just need to start, be willing to be wrong a hundred times, and keep going. JARVIS 2.0 is proof of that.

Back to Portfolio See MARS Robot →