How I built a micro-SaaS called Audiofern Home

How I built a micro-SaaS called Audiofern

5 min read
On this page

Last year at work, we started a book club for the SRE team where our goal at the time was to read a book a week and then come together and discuss as a team interesting things we’d learned.

The book we initially chose was called Engineering a Safer World by Nancy Leveson, a book which introduces a new systems-based approach (STAMP) for safety engineering, moving beyond traditional methods to address complex, software-driven systems, focusing on interactions and control rather than just component failures, and applying modern systems thinking to design safer, more reliable sociotechnical systems

As a vision impaired person (which some of my readers might know), I needed a way to “read” this book efficiently, there was no way for me to read this book the way everyone else would have, visually, as that would have taken me too long.

I need a way to be able to listen to the book and use my auditory senses so I could be prepared for the upcoming discussions. I knew that reading it the old fashioned way, which is impossible for me, would take far too long.

I came across a web service called Speechify that looked like it would do the trick and be able to take the book in PDF form and read the book to my out loud. The web interface also had a nice feature that highlighted the words as it read the book to you.

I used this service quite successfully to read through the first few chapters of the book, using the web application at every moment I had after work trying to get through each chapter as quickly as everyone else was so I could keep up.

This worked quite well, until the free trial ran out. My options to continue were sadly and frustratingly very limited. Speechify at the time had only one option, which was to pay $20/month (USD), but the problem was the free trial and forced upgrade option only let you pay for the entire year up-front! :O

To my absolute surprise there seemed to be no way for me to either:

  1. Pay as you go – Paying only for what I needed

or

  1. Paying just the monthly subscription that their pricing model allows.

Weirdly the only way for me to continue was to pay some ~$300 USD for the full year just for me to continue to read one lousy book! I was shocked, angry and disappointed.

I tried to contact their sales/support team to discuss how I might be able to continue to read my book and what options or flexibility they could afford to help me out with my plight.

To my absolute dismay, their staff at the time seemed unwilling or unable to help in any way so I decided to cancel my account. Thankfully they did. At the time I remember that as the free trial offer period was coming up, I was promptly contacted by their systems with amazing 50% discounts, but this still only applied to the entire year’s subscription! So I still had to pay over ~$150 USD to continue to use the service.

In the end I said to Speechify, I’ll just build my own thing™ using Google Voice or ChatGPT.

This was all back in March of 2025. Nearly a year later, here we are, I’ve built my own service called Audiofern that does exactly what I want. It lets you take any book, paper or article (currently only supporting PDFs right now, but more coming soon™), and uses OpenAI’s API(s) to extract text from the input as accurately as it can, and then OpenAI’s voice model(s) to convert that into nicely spoken narrated audio. The service (Audiofern) then stitches together the audio into a full audiobook with chapters and metadata, that you can listen to with the built-in web player or download to your device for offline listening.

Check it out at: https://audiofern.com

I build Audiofern in the course of two weekends, the first weekend was spent building and testing the web service, after first prototyping a quick command-line tool that proved out the concept. I spent hours testing and making sure everything works “just so” and my wife even helped me too! The next full weekend was spent productizing the service and getting it ready for launch.

I’ve done things like this before, built a thing on a weekend, but this time I used the help of “AI”. You can read more about my first forays into using AI Coding Agents in my post AI, Agents, and ‘Vibe Coding’ — Building Gatherly in Days, Not Weeks.

I won’t go into why I used “AI to help me build this service, but needless to say that the use of Claude and Codex has enabled me (as one person) to build a full e2e web service complete with payments, marketing, documentation and high quality in just a couple of weekends.

To date I’ve now read the following books/papers:

I feel like I’m a new person, a learned man! :-D Reading is made fun again!

#micro#saas #audio #tts #audiobook #audiofern