A smart, browser-native voice assistant powered by Gemini 2.5 Flash, packed with modular tools and a beautiful glassmorphism interface.
MAX uses advanced Natural Language Processing to figure out what you want and dynamically selects the right tool for the job.
| Feature | Description |
|---|---|
| 🎙 Native Voice Input | Push-to-talk microphone and hands-free "Hey Max" wake word via modern Browser Web Speech API (Zero complex C++ pyaudio dependencies!). |
| 🔊 Natural TTS | Text-To-Speech read aloud seamlessly via browser-native natural voices. |
| 🤖 Gemini 2.5 Flash AI | Highly intelligent conversational agent powered by Google's latest Gemini architecture. |
| Searches YouTube and generates a horizontal, highly interactive scrollable gallery. Clicking a card instantly opens a built-in Modal Iframe player to watch the video without leaving the chat! | |
| 🔍 Intelligent Web Search | Scrapes the web and routes the raw text back through Gemini for summarizing, so you get polished, conversational answers instead of link dumps. |
| 📧 Multi-User Email | Send emails directly from the chat. The UI features a slick modal allowing anyone using the app to securely enter their own SMTP credentials. |
| 🌤 Live Weather | Hooked up to OpenWeatherMap for instant, accurate forecasts anywhere. |
| 📈 Live Stocks | Hooked up to Yahoo Finance (yFinance) for instant stock and crypto tickers. |
- Glassmorphism Design: Dark mode gradients, animated background orbs, blurring, and stunning visuals via
style.css. - Quick Action Chips: Tap the suggested bubbles on the home screen to instantly execute commands without typing.
- Dynamic Modals: Smooth pop-ups for composing Emails, entering Stock/City names, and watching YouTube videos.
Because MAX now utilizes browser-native speech decoding, installation is incredibly fast and avoids the notoriously buggy PyAudio.
pip install -r requirements.txtCreate a .env file in the main folder (you can copy .env.example as a template).
GEMINI_API_KEY=your_gemini_key # Get from https://aistudio.google.com
WEATHER_API_KEY=your_owm_key # Get from https://openweathermap.org/api
# (Optional) Default Server Email Config:
EMAIL_ADDRESS=your_gmail@gmail.com
EMAIL_PASSWORD=your_app_password # Use a Google App Password, NOT your real passwordNote: If users do not want to use the default .env server email, the front-end Email Modal allows them to inject their own temporary credentials when sending an email!
python app.pyGo to http://localhost:5000 in your web browser.
(Note: To use the Microphone, modern browsers require that you run via localhost or serve the application securely over HTTPS).
MAX Bot is pre-configured for cloud deployment!
We use Gunicorn as our production WSGI server. A Procfile is already included in the repo.
- Push your repository to GitHub.
- Link the repository to a free host like Render.com or Heroku.
- Set your Build Command to:
pip install -r requirements.txt - Make sure to input your
GEMINI_API_KEYandWEATHER_API_KEYinto your cloud provider's Environment Variables panel. - Deploy safely over HTTPS!
- "Hey Max, what's the weather like in Tokyo right now?"
- "Get me the stock price for Nvidia (NVDA)"
- "Play lo-fi study girl music on YouTube" (Watch the interactive gallery pop up!)
- "Search the web for what happened to the Mars Opportunity Rover" (Watch Gemini summarize the raw search!)
- "Send an email to my friend"
- Backend Logic: Python 3.10+, Flask, Gunicorn
- AI Engine: Google
gemini-2.5-flash - Tooling:
yFinance,duckduckgo_search,smtplib, nativeurllibscraping module. - Frontend & State: Vanilla HTML/CSS/JS (Zero framework bloat), Web Speech API.