# Context-Aware Multimodal Assistant

## Overview
This project builds a multimodal assistant that helps users manage cognitive load by detecting stress from voice recordings and facial images. Based on the stress level, it simplifies or rephrases user tasks or messages to make them easier to understand.

## Features
- Detects stress from voice and face inputs (placeholder logic, easy to replace).
- Simplifies text input using the `facebook/bart-large-cnn` model from Hugging Face.
- Interactive UI built with Gradio for easy testing and deployment.

## How to Use
1. Upload a voice recording (.wav) and a face image.
2. Enter the task or message you want help with.
3. The assistant detects your stress level and simplifies your input accordingly.

## Technologies
- Hugging Face Transformers (`facebook/bart-large-cnn`)
- Gradio for the user interface
- Torch and other libraries for processing

## Future Work
- Integrate real stress detection models for voice and facial expressions.
- Add real-time input support.
- Extend functionality for calendar and email summarization.

---

Feel free to explore and contribute!