Transcribes video/audio/url to subtitles.
Project description
video2sub
video2sub is a command-line tool that transcribes video/audio/url to subtitles. It supports various subtitle formats such as txt, srt, vtt, tsv, and json. The tool uses speech recognition models from whisper to perform transcription.
Requirements
Python 3.8+
ffmpeg
Installation
Install ffmpeg.
(Optional) Install CUDA & cuDNN. You can either install them system wide or create a conda environment using environment.yml
.
Install pipx.
Install video2sub using pipx:
pipx install video2sub --include-deps
(Optional) Copy config.vide2sub.example.yml
to ~/.config/
(win %UserProfile%\.config
) then rename it to config.video2sub.yml
. Edit it to your prefer default value.
Usage
video2sub [video_url] [--lang] [--model] [--proxy] [--output] [--verbose] [--force]
Argument | Description |
---|---|
video_url | The URL of the video or path of the video/audio file. |
--lang | The language code of the audio file to be transcribed. Default: "en". |
--model | The name/path to the speech recognition model to be used for transcription. Default: "small". |
--proxy | Proxy to be used for downloading the video from url. |
--output | Output format: txt, vtt, srt, tsv, json. Default: "txt". |
--verbose | Show output. |
--force | Force to retrieve then transcribe video regardless of existing audio or subtitle file. |
Note: If the --lang or --model arguments are not specified, the tool will use the default values which can be set by the user in the configuration file ~/.config/video2txt.yml
.
Examples
-
Transcribe a video from a URL using default config:
video2sub https://www.youtube.com/watch?v=dQw4w9WgXcQ
-
Transcribe a video from a English spoken video file using medium model and save it as a .vtt subtitle:
video2sub my_video.mp4 --lang en --model medium --output vtt
-
Transcribe an audio file and save it as a .srt subtitle:
video2sub my_audio.mp3 --output srt
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for video2sub-0.1.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | d579886c256aa82014cc1e4ec0efbe0b0d1f5f3eade817964af38b7b8e744b9d |
|
MD5 | 8a254d7465e1159abecfed6212f1cb11 |
|
BLAKE2b-256 | f82e82c1f0c1ab0b2f1e973cc74da351b6f427bc55200a98e37623bc4534cb1e |