Scamurai is an AI-powered system that detects deepfake audio from voicemails or recordings. It uses a deep learning model (RawNet2) to analyze raw audio and determine whether a voice is real or AI-generated.

Due to the rise of AI-generated voices, scams have become a lot more convincing and harder to detect. It is especially harder for the elderly who are targeted through phone calls and voicemails. We wanted to build a tool that would help people verify whether a voice is real or fake, before they act on it.

• Upload or record audio
• Detect deepfake vs real voice
• Fast inference using trained model
• Designed for mobile app integration (SwiftUI)

We built Scamurai using RawNet2. It is a deep learning model that processes raw audio directly.

This is how the code works:

• Upload an audio file (voicemail or recording)
• Preprocessing of the file (resampling, trimming, normalization)
• Deep learning model: Audio -> Sinc Filters -> Residual CNN -> GRU -> Prediction
• Outputs Real or Fake with a confidence score

We trained the model using a gender-based labeled dataset of real and spoofed audio, and integrated the system with a mobile app where users can upload audio for analysis.

• PyTorch – Deep learning framework
• RawNet2 – Deepfake audio detection model
• FastAPI – Backend API
• SwiftUI – iOS frontend

Audio Processing: NumPy, SciPy, SoundFile
Languages: Python, Swift

pip install -r requirements.txt

python server.py

Server will run at:

http://localhost:8000

• Go to settings in your iPhone and turn on Developer Mode
• Open the application in XCode
• Connect your iPhone through a USB cable and select it in Device Selector
• Go to Signing & Capabilities and select your Apple ID under Team
• Click the Run button

POST /analyze

Audio file (.wav)

Example (Python request):

import requests

files = {"file": open("sample.wav", "rb")}
response = requests.post("http://localhost:8000/analyze", files=files)

print(response.json())

{
  "label": "bonafide",
  "confidence": 0.93
}

python infer.py path/to/audio.wav

In the future, we plan to:

• Improve accuracy even more with larger and more diverse datasets
• Enable real-time call detection
• Improve accessibility for elderly users

Scamurai helps make phone calls safer by giving users a simple way to verify authenticity in an age of AI-generated content.

Spot the fake. Trust the real.