jaredgth/audible-bookmark-transcriber

A tool to automatically transcribe Audible bookmarks for easier export

Audible bookmark transcriber

Introduction

This is a small tool for Audible users to easily extract their bookmarks from the audio books as text. By default, it uses the state-of-the-art speech-to-text AI model called Whisper by OpenAI.

Important

This tool is only meant to help users download and use the Audible books they actually own.

Limitations

• It only supports English books for now.
• It has only been tested on MacOS. Linux will probably work, but Windows is unlikely.

User guide

Follow these steps to get started.

Clone this repo.

Install the necessary dependencies.

1. ffmpeg (if you need to install brew, follow the instructions here)

brew install ffmpeg

2. jq

brew install jq

3. The rest are Python dependencies that can be installed by running the following command. Consider using an environment management tool such as Conda.

pip install -r requirements.txt

Log in to your Audible account and provide an OpenAI API key

Run:

python main.py login

The command will ask for a few information:

• Your Audible user name (for some users, this is your Amazon user name)
• Your Audible password. This will only be used for authenticating to your Audible account.
• The counry/region of your Audible account.
• Your OpenAI API key. If you don't have one, follow the steps here.

Get a list of your Audible books

Run:

python main.py list

If everything works as expected, you should see a list of Audible books (ASIN + title) printed.

Extract bookmarks and transcribe

Run:

python main.py transcribe_bookmarks <ASIN>

where <ASIN> is the book ID you want to use.

Wait for the code to run

It will take a while. Basically the code performs the following steps:

1. Download the original audio book file, remove the DRM using your Audible user credential, and save it as a .mp4 file.

Important

Anti-Piracy Notice

Please only use these tools for gaining full access to your own ebooks for archiving/conversion/convenience. De-drmed ebooks should not be uploaded to open servers, torrents, or other methods of mass distribution. No help will be given to people doing such things. Authors, retailers and publishers all need to make a living, so that they can continue to produce books for us to read. Don’t be a parasite.

The converted .mp4 file is only meant for your personal use. DO NOT distribute! After the code finish running, you should delete this file to avoid leaks.

2. Extract the bookmarks you created for this book, specifically, the start and end time of each clip.
3. Extract each clip as separate audio files.
4. Call the Whisper API to transcribe the audio clips.

Completed.

The transcribed notes will be saved in this file: data/audio/<ASIN>/transcribed_notes.json.