Silero VAD with FastAPI

This repository contains a robust implementation of Silero Voice Activity Detection (VAD) integrated with FastAPI, providing an efficient and scalable solution for real-time audio processing.

Features

Silero VAD Integration: Utilizes the state-of-the-art Silero VAD for accurate voice activity detection. FastAPI Framework: Leverages FastAPI for a high-performance, asynchronous web framework that's easy to use and ideal for building APIs.

Getting Started

Prerequisites

• Python 3.8+
• FastAPI
• Silero VAD

Installation

1. Clone repository

git clone https://github.com/Bronnzy/SileroVAD_FastApi.git

2. Install the required packages:

pip install -r requirements.txt

Running the application

1. Start the FastAPI server

uvicorn vad_check_api:app --host 0.0.0.0 --port 4499

2. Access the API documentation: Open your browser and navigate to http://localhost:4499/docs to explore the interactive API documentation provided by Swagger UI.

API Endpoints

• POST /vad_check/: Accepts audio/video buffer(blob) and indicates if speech detected and returns segments of detected speech.

  • Request
    • Content-Type: multipart/form-data
    • Body: An audio file (e.g., WAV format) uploaded using a form field named 'file'.
  • Response
    • Content-Type: application/json
    • Body: A JSON object containing:
    • has_voice (bool): Indicates whether any speech was detected in the audio.
    • speech_timestamps (list): A list of objects representing the start and end timestamps (in seconds) of detected speech segments..

• POST /vad_check_file/: Accepts an audio/video file and indicates if speech detected and returns segments of detected speech.

  • Request
    • Content-Type: application/octet-stream
    • Body: Raw binary audio data
  • Response
    • Content-Type: application/json
    • Body: A JSON object containing:
    • has_voice (bool): Indicates whether any speech was detected in the audio.
    • speech_timestamps (list): A list of objects representing the start and end timestamps (in seconds) of detected speech segments