Fake Voice Detection

Introduction

The "Fake Voice Detection" project aims to develop a mobile application capable of identifying fake voices during live phone calls. This application will enhance user security by detecting and alerting users to potential voice manipulations, such as mimicry or AI-generated voices, and providing location information of the caller.

System Overview

The Fake Voice Detection system is designed to operate on mobile devices, providing real-time analysis of voice data during phone calls. It will identify whether the voice is real, mimicked, or AI-generated. Upon detecting a fake voice, the system will alert the user through vibrations and notifications, and it will also attempt to identify the caller's location.

Functional Requirements

• As a User, I should be able to receive real-time analysis of voice data during phone calls to detect real, mimic, or AI-generated voices.
• As a User, I should be alerted through vibrations and notifications when a fake voice is detected.
• As a User, I should be able to see the location of the caller when a fake voice is detected.

User Personas

• General User: Individuals using the app to ensure the authenticity of phone calls and protect themselves from voice spoofing.
• Student: Users like Aakash, who are developing projects or conducting research on voice detection technologies.

Visuals Colors and Theme

• primary: #1A237E (Deep Indigo)
• primary_light: #534BAE (Light Indigo)
• secondary: #FF6F61 (Coral)
• accent: #FFD700 (Gold)
• highlight: #FFB300 (Amber)
• bg: #F5F5F5 (Light Gray)
• surface: rgba(255, 255, 255, 0.9) (White with slight transparency)
• text: #212121 (Dark Gray)
• text_muted: #757575 (Muted Gray)
• border: rgba(33, 33, 33, 0.1) (Subtle Gray)

Signature Design Concept

Imagine a dynamic, interactive home screen that resembles a soundwave visualizer. As users speak or receive calls, the visualizer animates in real-time, reflecting the voice's authenticity. When a fake voice is detected, the visualizer shifts colors dramatically, from calming blues to alerting reds, and vibrates subtly to draw attention. Users can interact with the visualizer by tapping to see detailed analysis, including the type of voice detected and the caller's location. This concept will be implemented using framer-motion for smooth animations and react-icons for intuitive iconography.

Interaction Model & Motion Direction

The landing page will utilize a "parallax" interaction model, providing a layered depth effect as users scroll through the app's features. This will create an engaging storytelling experience, with decorative layers moving at different speeds to enhance the visual appeal. For internal pages, a "static" model will be used to maintain clarity and focus on functionality.

Non-Functional Requirements

• The app should perform real-time voice analysis with minimal latency.
• It must operate efficiently on mobile devices without significant battery drain.
• The user interface should be intuitive and accessible to a broad audience.

Tech Stack

• Frontend: React Native for mobile app development.
• Backend: Python with FastAPI for handling voice analysis.
• Database: MongoDB for storing user data and analysis results.
• AI Models: GPT 5.4 for user-friendly responses, Claude 4.6 Opas for academic or coding work.
• AI Tools: Litellm for LLM Routing, Langchain.
• Local Orchestration: Docker, docker-compose.
• Server-side Orchestration: Kubernetes.

Assumptions and Constraints

• The app assumes access to the device's microphone and location services.
• It is constrained by the processing power and battery life of mobile devices.
• The accuracy of voice detection may vary based on environmental noise and call quality.

Glossary

• AI-generated Voice: A voice created using artificial intelligence technologies.
• Mimic Voice: A voice that imitates another person's voice.
• Real-time Analysis: The process of analyzing data as it is received, without delay.
• Parallax: A visual effect where background images move slower than foreground images, creating a sense of depth.

This document outlines the core requirements and design concepts for the Fake Voice Detection project, ensuring a clear path forward for development and implementation.