AI_Voice_Agent Project

[Saloni90sharma]

Open Source Hackathon 2026 Project Submission

Participant Details

Full Name:
Saloni Sharma

GitHub Username:
Saloni90Sharma

Team Name:

College/University:
GLA University Mathura

---

Project Details

Project Title:
AI Voice Agent

Project Description:

ARIA – AI Voice Agent

ARIA is a modern AI-powered voice assistant that enables natural, real-time voice conversations between users and artificial intelligence. Built with FastAPI, Google Gemini 1.5 Flash, and Web Speech APIs, the application allows users to speak directly to the assistant, receive intelligent context-aware responses, and hear those responses through natural voice synthesis.

The project features session-based conversation memory, real-time speech recognition, text-to-speech capabilities, and a responsive glassmorphism-inspired user interface. Designed with simplicity and performance in mind, ARIA uses a lightweight frontend built with HTML, CSS, and JavaScript, making it easy to deploy and maintain.

Key Features

• 🎤 Real-time Voice Input using Web Speech API
• 🧠 Intelligent Responses powered by Google Gemini 1.5 Flash
• 💬 Session-based Conversation Memory
• 🔊 Natural Voice Output with SpeechSynthesis API
• ⚡ FastAPI Backend with REST APIs
• 🎨 Modern Responsive UI with Dark Theme
• 🔄 Automatic Speech Recognition Recovery
• 📱 Mobile-Friendly Design
• 🚀 Ready for Deployment on Render and GitHub

Tech Stack

• Frontend: HTML5, CSS3, JavaScript
• Backend: Python, FastAPI, Uvicorn
• AI Model: Google Gemini 1.5 Flash
• Speech Recognition: Web Speech API
• Text-to-Speech: SpeechSynthesis API
• Deployment: Render, GitHub

Use Cases

• Virtual AI Assistant
• Customer Support Automation
• Educational Learning Assistant
• Productivity and Information Retrieval
• Voice-Based Human-Computer Interaction

ARIA demonstrates how Generative AI, voice technologies, and modern web development can be combined to create an intelligent, scalable, and user-friendly conversational assistant.

Tech Stack Used:
Frontend: HTML5, CSS3, JavaScript
Backend: Python, FastAPI, Uvicorn
AI Model: Google Gemini 1.5 Flash
Speech Recognition: Web Speech API
Text-to-Speech: SpeechSynthesis API
Deployment: Render, GitHub

GitHub Repository Link:

Live Demo Link:

Presentation / Demo Video Link:

---

Open Source Readiness

• [✓] My project is public on GitHub
• [✓] My repository has a proper README.md
• [✓] I have added setup/installation instructions
• [✓] I have added screenshots/demo where possible
• [✓] I have added a license file
• [✓] My project is original and built/updated during the hackathon period

---

Memori Labs Sponsor Task

Please complete these before submitting:

• [✓] I have starred the Memori Labs GitHub repository
  https://github.com/MemoriLabs/Memori

• [✓] I have followed Memori Labs on LinkedIn
  https://www.linkedin.com/company/memorilabs/

• [✓] I have followed Memori Labs on X
  https://x.com/memorilab

• [✓] I have checked Memori Labs social links
  https://linktr.ee/memorilabs

---

ID Card Verification

• [✓] I have generated my ID card from https://oshack.xyz
• [✓] If my ID was not verified, I completed the mandatory verification/giveaway form and tried again

---