User Guide

The audioarxiv command-line tool allows you to fetch research papers from arXiv and read them aloud using a text-to-speech engine.

Basic Usage

$ audioarxiv --id 1602.03837

This will fetch the arXiv paper with ID 1602.03837, process its content, and read it aloud.

To save the audio to a file instead:

$ audioarxiv --id 1602.03837 --output audio.mp3

Reminder: It takes a long time to process a paper into an audio file.

Command-Line Options

Command-line Options for audioarxiv

Option	Description
-h, --help	Show the help message and exit.
--id ID	arXiv paper ID.
--output OUTPUT	Output to audio file if provided.
--rate RATE	Number of words per minute between 50 and 500.
--volume VOLUME	Volume between 0 and 1.
--voice VOICE	Voice to use for text-to-speech.
--pause-seconds PAUSE_SECONDS	Duration of pause between sentences in seconds.
--page-size PAGE_SIZE	Maximum number of results fetched in a single API request.
--delay-seconds DELAY_SECONDS	Number of seconds to wait between API requests.
--num-retries NUM_RETRIES	Number of times to retry a failing API request before raising an exception.
--list-voices	List the available voices.

List Available Voices

To explore which voices are supported by your system:

$ audioarxiv --list-voices

Configuration File

All command-line options can be saved in a configuration file for reuse. When the tool is run, it generates or loads settings from a config file stored locally. The path to this file is printed in the logs when you run the tool with --id.

This allows you to define and reuse your preferred settings (e.g. voice, volume, rate) without needing to type them every time.