Foreign Whispers - YouTube Video Downloader

Foreign Whispers is a Django web application that allows users to download YouTube videos with captions and transcripts in multiple languages along with a bunch of AI models to translate the generated text to the user-desired language. The application uses the pytube library to download videos and captions, and the whisper library to generate transcripts. The application also uses the Mozilla TTS library to generate audio files from the transcripts and a mix of librosa, soundfile and moviepy to embed the audio file into the video.

YouTube Demo Link

https://youtu.be/3plIfwKIbFQ

Example outputs of the application

1. English to French: Drive Link
2. English to German: Drive Link

Installation

The installation and project setup can be done in two ways:

Installing locally (Given you have python in your local machine)

1. Clone the repository to your local machine.
2. Create a Python virtual environment and activate it.
3. Install project dependencies using pip install -r requirements.txt.
4. It also requires rubberband to be installed on your system.
5. Run Django migrations: python manage.py migrate.
6. Start the development server: python manage.py runserver.
7. Access the application in your web browser at http://localhost:8000/video_downloader/download.

Booting up the project using Docker, which only needs docker installed on your machine.

1. Open the root of the project (where the Dockerfile is installed)
2. Execute the command: docker-compose up --build -d
3. It will take about 300-900s for the initial project setup depending on your network configuration.
4. It launches a docker image on successful build containing the project which can be access in your web browser at http://localhost:8000/video_downloader/download.

Usage

1. Visit the homepage and provide a YouTube video URL.
2. Click the "Download" button to initiate the download process.
3. The video, captions (if available), and transcript (if necessary) will be saved in the media/ directory.

Project Screenshots

Milestone 1

Milestone 2

Milestone 3

Milestone 4

The application will return an audio file within the media directory of the selected translation in a .wav format.

Milestone 5

Milestone 6

Team Members

• Malhar Patel (mkp6112@nyu.edu)
• Ruben Garcia (rg4352@nyu.edu)

License

This project is licensed under the MIT License.

Contact

For questions or support, please contact Malhar Patel.