Introduction

The Gladia CLI is a command-line interface tool that allows developers to transcribe audio files or URLs using the Gladia API. It supports several features such as noise reduction, speaker diarization, direct translation, text emotion recognition, and summarization.

Installation

The Gladia CLI can be installed directly from the GitHub repository or built from source using pipenv.

Direct Install

To install the Gladia CLI directly from the GitHub repository, use the following commands:

Linux

wget <https://github.com/gladiaio/gladia-cli/raw/main/dist/linux_x64_gladia> &&  
mv linux_x64_gladia gladia &&  
chmod +x gladia

MacOS ARM

wget <https://github.com/gladiaio/gladia-cli/raw/main/dist/macos_arm64_gladia> &&  
mv macos_arm64_gladia gladia &&  
chmod +x gladia

Build from Source

To build the Gladia CLI from the source, use the following commands:

pipenv shell  
pip install -r requirements.txt  
./build.sh  
Usage  
The Gladia CLI can be used by running the gladia command followed by the necessary options.

Options  
The following are the available options for the Gladia CLI:

\--audio-url: The URL of the audio file to be transcribed.  
--audio-file: The path to the audio file to be transcribed.  
--language-behaviour: Determines how to handle multi-language audio.  
--language: The language spoken in the audio file.  
--transcription-hint: A hint to the transcription model. You can pass names, topics, custom vocabulary, etc.  
--noise-reduction: Applies noise reduction to the audio.  
--diarization: Performs speaker diarization.  
--diarization-max-speakers: Determines the maximum number of speakers to be detected.  
--direct-translate: Activates direct translation to the specified language.  
--direct-translate-language: The language to which to translate the transcription. Requires the direct translation option to be activated using --direct-translate.  
--text-emotion: Activates text emotion recognition.  
--summarization: Activates summarization.  
--output-format: The format in which to return the transcription results. Possible values are table, json, text, srt, vtt, plain.  
--gladia-key: The API key for Gladia. Get it at <https://app.gladia.io/account>  
--save-gladia-key: Saves the API key to a configuration file.  
--help: Shows the available options.

Authentication

To use the Gladia API, developers must obtain an API key from https://app.gladia.io/account.

The API key can be saved to a configuration file using the --save-gladia-key option.

Basic Example

The following is an example of using the Gladia CLI to transcribe an audio file:

gladia --audio-url http://files.gladia.io/example/audio-transcription/split_infinity.wav

Transcribing audio file...  
Transcript

 time_begin  time_end  probability  language  speaker        transcription  
 0.09        2.07      0.49         en        not_activated  Split infinity  
 2.13        5.19      0.65         en        not_activated  in a time when less is more