How chandanswain9/personalized-ai works
Project type and purpose
• This is a hybrid full-stack “desktop” web-view application (using Python + Eel to host an HTML/JS frontend) that functions as a voice-activated AI assistant.
• Purpose: capture microphone input, run speech recognition → parse commands → invoke backend features (e.g., open apps, search YouTube, send WhatsApp) → return responses via text + animation + speech synthesis.
File structure
• Top level:
• run.py / main.py / test.py – entrypoints and smoke tests
• contacts.csv – simple datastore for WhatsApp contacts
• engine/ – Python backend modules
• engine/auth/ – face-recognition (haar cascade, trainer, sample code) for optional user authentication
• engine/command.py – central command dispatcher
• engine/features.py – implementations for each assistant feature (open apps, search, messaging)
• engine/db.py – lightweight data access (contacts)
• engine/config.py – configuration parameters (API keys, hotword settings)
• engine/helper.py – shared utilities
• www/ – frontend served by Eel
• index.html + style.css – UI skeleton and styles
• main.js / controller.js / script.js – frontend logic, event handlers, Eel bindings
• www/assets/ – CSS/JS libraries (Textillate, Animate.css) and media (audio wave animation)
README for additional insights
• Tech stack highlights:
• Frontend: HTML/CSS/JS + jQuery + Textillate.js + Animate.css
• Backend: Python3 + pyttsx3 (TTS) + speech_recognition + eel + gevent + whisper (optional)
• Key features: hotword detection, animated chat UI, voice I/O, app integration
Elements
– Blue rectangles for internal modules