Saathi - Your Digital Companion 🤝

Saathi is an AI-powered assistant designed to help older adults navigate the digital world with confidence, providing simple, step-by-step guidance for everyday online tasks in both English and Hindi.

Demo Video

🎯 The Problem

Many older adults find it challenging to keep up with modern technology. Essential tasks like paying bills online, sending an email, or using a messaging app can be confusing and intimidating. Saathi was built to be a patient, friendly companion that bridges this gap.

✨ Features

• Conversational Text Chat: A fully functional chat interface that remembers the conversation's context and renders formatted Markdown responses for clarity.
• Advanced Voice Assistant: A complete voice-to-voice pipeline. Users can speak their query in English or Hindi and receive step-by-step guidance spoken back by the browser's native voice.
• Hands-Free Voice Commands: Once the steps are displayed, users can navigate them using voice commands like "next," "previous," and "repeat" in both English and Hindi, providing a true hands-free experience.
• Purpose-Driven AI: The assistant is specifically engineered with prompt guard rails. It is an expert on digital tasks and will politely decline off-topic questions (e.g., general knowledge, trivia) to stay focused on its mission.
• Smart App Recommendations: Instead of using technical jargon like "email client," Saathi is instructed to recommend common, popular apps like "Gmail" or "Paytm" to make its advice more actionable.
• Secure Authentication & History: Built with Google Firebase for secure and easy user login. All chat conversations are saved to a persistent database and are accessible via a slide-out history panel.

📸 Screenshots

🛠️ Tech Stack

• Frontend: React, Material-UI (MUI), Axios
• Backend: Java, Spring Boot, Spring Security, PostgreSQL, Maven
• AI/ML:
  • Text Generation: Cerebras API (llama-4-scout-17b-16e-instruct)
  • Speech-to-Text: Groq API (whisper-large-v3)
• Authentication: Google Firebase

🚀 Getting Started

Prerequisites

• Java 17+
• Node.js & npm
• A running PostgreSQL instance
• API keys for Firebase, Groq, and Cerebras

Backend Setup

1. Clone the repository.
2. Navigate to the Saathi-Backend directory.
3. Set up your environment variables in src/main/resources/application.properties with your database credentials and API keys.
4. Run the application: ./mvnw spring-boot:run

Frontend Setup

1. Navigate to the Saathi-Frontend directory.
2. Install dependencies: npm install
3. Create a .env.local file and add your Firebase configuration keys.
4. Run the application: npm run dev

🔮 Future Features

• High-Quality Multilingual TTS: Integrating a cloud TTS service for more natural-sounding voice responses.
• Screen Sharing Assistance: A feature to allow Saathi to guide users visually on their shared screen.
• Saving and Favoriting Chats: Allowing users to save and easily access important guides.