Back

⚙️ DIY Audiobooks: Turn Your eBooks into Natural-Sounding Audio with ebook2audiobook & Docker

Ever wished you could listen to that niche ebook, academic paper, or long article that doesn't have a commercial audiobook version? Inspired by services like "Read To Me" which turn documents into high-quality audio, you might be looking for a way to do this yourself.

Good news! With the open-source project ebook2audiobook and a bit of setup using Docker, you can transform your texts into natural-sounding audiobooks, complete with features like chapter navigation. This guide will walk you through the process, even if you're new to Docker.

Why This DIY Approach?

• Local Control: Process your documents on your own machine.
• Customization: Choose from various text-to-speech (TTS) engines and voices.
• Open Source: Leverage a community-driven project.

What You'll Need:

1. A Computer (Windows, Mac, or Linux): The process is similar across platforms.
2. Docker Desktop: This is essential for running ebook2audiobook easily. Download Docker Desktop here.
3. Your eBook/Document: PDF and EPUB are well-supported.
4. A Bit of Patience: Especially for the initial setup and longer books.

Step-by-Step Guide to Your First DIY Audiobook

Step 1: Install Docker Desktop

If you haven't already, download Docker Desktop from the link above and install it on your computer. Docker allows us to run applications in isolated environments called containers, making setup much simpler.

Step 2: Get the ebook2audiobook Docker Command

The ebook2audiobook project lives on GitHub. The project's README page (usually the main page you land on) contains the necessary commands.

• Go to the ebook2audiobook GitHub repository (or search for "ebook2audiobook" on GitHub).
• Scroll down to the "Docker" section, specifically "Running the Docker Container".
• You'll see two primary commands:
  • For CPU Only (If you don't have an NVIDIA GPU or are unsure):

    docker run --pull always --rm -p 7860:7860 athomasson2/ebook2audiobook
    

  • For NVIDIA GPU Users (Faster processing):

    docker run --pull always --rm --gpus all -p 7860:7860 athomasson2/ebook2audiobook
    

  Copy the command that's appropriate for your system. Using an NVIDIA GPU will significantly speed up the conversion process.

Step 3: Run the Docker Command in Your Terminal

1. Open your computer's terminal application:
   • Windows: Search for "Terminal", "PowerShell" or "Command Prompt."
   • Mac: Search for "Terminal."
   • Linux: Open your preferred terminal.
2. Paste the Docker command you copied in Step 2 into the terminal and press Enter.
3. First-Time Run: The first time you run this, Docker will need to download the ebook2audiobook image/application. This might take a few minutes, depending on your internet speed. You'll see various download progress messages.
4. Success: Once the image is downloaded and the container starts, you should see messages indicating it's running, potentially ending with something like "v25.x.x Full_docker mode" or similar log output indicating the service is active on port 7860.

Step 4: Access the ebook2audiobook Web Interface

Once the Docker container is running, open your web browser (Chrome, Firefox, Edge, etc.) and go to the following address:

http://localhost:7860

This will open the Gradio web interface for ebook2audiobook.

Step 5: Configure and Convert Your eBook

You'll now see the interface for converting your document:

1. Upload Your File: Drag and drop your PDF or EPUB file into the "Select a File" or "Drop File Here" area.
2. Language: Select the language of your document (e.g., English).
3. TTS Engine: Choose a Text-to-Speech engine. "XTTS" is often a good default known for natural-sounding voices.
4. Fine Tuned Models (Voice Selection):
   • For XTTS, "internal" is a good starting point.
   • You can explore other pre-trained voices listed if you want to experiment.
5. Processor Unit:
   • If you used the GPU Docker command and have an NVIDIA GPU, select GPU.
   • Otherwise, select CPU.
6. Output Format:
   • M4B: Highly recommended for audiobooks. This format supports chapters (making navigation easy) and resume playback features in compatible players.
   • MP3: A standard, widely compatible audio format.
7. Start Processing: Once you've configured your options, click the main processing button.

Step 6: Download and Listen

• Processing Time: The conversion can take some time. As mentioned in the video:
  • With a GPU, it might be around 2 minutes per page.
  • With a CPU, it could be 18 minutes per page or more.
  • This means a 250-page book could take around 8 hours on a GPU, or a hefty 75 hours on a CPU! Be patient, especially with longer books or if using CPU only.
• Download: Once processing is complete, a download link or the file itself (often named index.m4b or index.mp3) will appear in the "Audiobook" section of the interface. Click to download it.
• Listen: You can now play your audiobook using any compatible audio player.
  • For M4B files, players like VLC Media Player, Apple Books, or dedicated audiobook apps will allow you to see and navigate by chapters.

Important Considerations:

• Experiment with Voices: The default voices are good, but feel free to try different "Fine Tuned Models" to find one you prefer for your content.
• Processing Power: A dedicated NVIDIA GPU makes a huge difference in speed. If you plan to convert many long books, it's a significant advantage.
• Document Quality: The cleaner your source document, the better the audio. ebook2audiobook does some preprocessing, but very messy PDFs might still have quirks.

Conclusion

You've now successfully set up a powerful local ebook-to-audiobook converter! While the initial setup might seem a bit technical, Docker simplifies the process significantly. Enjoy listening to your documents in a whole new way, with the exact voices and formats you prefer. Enjoy!