Detect Page
Storage
Generation2
Home Page
Generation1

FoiceDetect: AI-Powered Voice Detection and Classification

Project Overview

We developed Fake Voice Detect, a platform leveraging Nebius AI to identify fake voices (deepfakes) and protect users from scams. As scammers increasingly use AI-generated voices to impersonate friends and request money, our platform evaluates how fake a voice is, provides Nebius AI-driven cybersecurity solutions, and even suggests humorous responses to outsmart scammers. Additionally, we securely store information for future use, including reports, evidence, and identification, empowering users to stay safe and take action against fraud.

Tracks:

Cybersecurity Track (WithSandra) $325
Most Technically Impressive - Modal (Series A company) $2,000 credits grand prize
Nethenoob spinning cat ($65 cash)
Nebius Exploratory Use of LLMs and Vision Models:
- 1st Place: $2,000 credits
- 2nd Place: $1,000 credits
- 3rd Place: $500 credits
$1625 CASH (STATSIG GRAND PRIZE)
Magic loops $325 CASH
Beginner tracks (20 winners) ($10 each credits)

🚀 Technologies Stack

Frontend

React.js
React Bootstrap
Axios
JavaScript ES6+

Backend

Python
Django
Machine Learning Libraries ### Backend Machine Learning Libraries
- NumPy
- Scikit-learn
- SVC (Support Vector Classification)
- train_test_split
- StandardScaler
- Librosa (Audio feature extraction)
- Joblib (Model serialization)
AssemblyAI Integration

Key Features

Voice Detection
Voice Classification
Real-time Audio Analysis
Machine Learning Model Training

🛠 Installation and Setup

Prerequisites

Python 3.8+
Node.js 14+
pip
git

Repository Clone

git clone https://github.com/dahomita/foicedetect.git
cd foicedetect

API Key Setup

Environment Configuration

Create a .env file in the project root directory
Add the following API keys to the .env file:

NEBIUS_API_KEY=eyJhbGciOiJIUzI1NiIsImtpZCI6IlV6SXJWd1h0dnprLVRvdzlLZWstc0M1akptWXBvX1VaVkxUZlpnMDRlOFUiLCJ0eXAiOiJKV1QifQ.eyJzdWIiOiJnb29nbGUtb2F1dGgyfDExMTY5ODA4MDkyNDAxNzkyMjcxOSIsInNjb3BlIjoib3BlbmlkIG9mZmxpbmVfYWNjZXNzIiwiaXNzIjoiYXBpX2tleV9pc3N1ZXIiLCJhdWQiOlsiaHR0cHM6Ly9uZWJpdXMtaW5mZXJlbmNlLmV1LmF1dGgwLmNvbS9hcGkvdjIvIl0sImV4cCI6MTg5NDY1NTAyMCwidXVpZCI6Ijc2NGNlOTc1LWI4ZjUtNDNkNi1hMTI1LTgzNDQwY2Y5MWVmYSIsIm5hbWUiOiJVbm5hbWVkIGtleSIsImV4cGlyZXNfYXQiOiIyMDMwLTAxLTE0VDIxOjAzOjQwKzAwMDAifQ.UXQkWATEndFFSgnZss62eQ9c7S64AfGLy2qOLJtDKQ8
ASSEMBLY_API_KEY=e98e3dcc13624e2c846c6de74a34809e

⚠️ Security Warning:

Please don't use our key too much

Backend Setup And Run

cd backend
python -m venv venv
venv\Scripts\activate
cd foicedetect_backend
pip install -r requirements.txt
pip install assemblyai
python manage.py makemigrations
python manage.py migrate
python manage.py runserver

Frontend Setup And Run

npm install
npm start

💡 Inspiration, Learning, and Challenges

What Inspired Us

We were inspired by personal experiences, as many of our friends and family members have been scammed through fake voice phone calls. Scammers impersonated their voices to deceive loved ones into sending money. This motivated us to create a platform that helps people like us by providing cybersecurity solutions to enhance not only their safety but also their knowledge.

What We Learned

How to use React for a dynamic frontend. How to use Django and Python for a flexible backend. How to implement User Authentication and Document Saving. Training Machine Learning models effectively. Integrating AI APIs for advanced functionality. Using Nebius to generate text and improve AI prompting. Seamlessly combining frontend and backend technologies.

Challenges Faced

Insufficient data for effective machine learning training. Managing secure storage and retrieval of user voice data. Generating accurate analysis and actionable solutions.