Quality Subs Generator

Effortlessly generate high-quality subtitles for your videos using the "quality-subs-generator" npm package. This powerful tool combines the capabilities of ffmpeg and the advanced Whisper model, providing a seamless solution for converting videos into various quality formats with precise subtitles.

Installation

Install the package easily using npm or Yarn:

Using npm

1npm install quality-subs-generator

Using Yarn

1yarn add quality-subs-generator

Usage

This package utilizes Docker for all processing, ensuring a hassle-free experience for users. You only need Docker; no need to deal with the tedious installation of components like the Whisper model, ffmpeg binaries, etc. The isolated environment also reduces the risk of system crashes.

During the initial run, the package will build a Docker image. This process includes heavy installations, such as ffmpeg, the OpenAI-Whisper model, transcription models, Node.js runtime, and more. Please note that it may take some time (approximately ~30 minutes) depending on internet speed and system resources, resulting in an image size of around ~10GB. Subsequent video processing will not rebuild the image; they will directly run a container.

Note: If you change modelName in whisperOptions, the image will rebuild. However, if you haven't deleted the Docker image created for building the required image, only the model downloading stage will run. In case you delete the image, the entire build process will restart. It's essential to note that after each change in modelName, new images will be created, but they won't consume significant space.

Example

1import startProcessing from 'quality-subs-generator';
2
3const processVideoWithSubtitles = async () => {
4    try {
5        await startProcessing({
6            inputVideoPath: `${process.cwd()}/input/video.mp4`,
7            outputDirectory: `${process.cwd()}/output`,
8            qualities: ['144'],
9            logs: true,
10            imageBuildOptions: { logs: true },
11            whisperOptions: { outputFormat: 'srt', modelName: 'tiny' }, // Default output format is srt, and the default model is 'tiny'
12        });
13        console.log('Service finished');
14    } catch (error) {
15        console.error(error);
16    }
17};
18
19processVideoWithSubtitles();

Dependencies

• dockerode: ^4.0.2
• random-word-slugs: ^0.1.7

Contributing

We welcome contributions! Feel free to contribute by raising issues or submitting pull requests. Please follow our contribution guidelines.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contact

For any inquiries or feedback, please reach out to Keshav Garg. Your input is valuable!