The Story of My Text-to-Speech App
Frequent readers of this blog know that I started writing an app to occupy myself through the dark winter months here. What started as mere dabble turned into full-blown obsession: during the day when the kids are in school, lunches were dreadful interruptions, and at night I dreamed about different ways of solving thorny problems. In November and December, I squeezed out every minute of time that doesn’t involve family life to work on the app.
In those two months, I had no idea whether the app would ever see the light of day. I used the text-to-speech engine that comes with iPad and iPhone. Apple deems the speech engine off-limits for third party apps, so my software rested on what’s essentially a hack. Apple would never allow it in the App Store.
After our Christmas and New Year trip to Oslo, I almost abandoned the project. Until I found an excellent text-to-speech engine built by another company, Acapela. I replaced the hack with their engine and all of a sudden, I had a viable product. Early February I submitted the app to Apple and a week later, it was approved and out in the wild. Save from a major mishap (a big bug which I quickly fixed), the app did much better than I expected in the beginning. During the first two days there were 100 downloads, which netted me a whopping $100. How did they find it? I didn’t do any promotion.
The feature set of the initial release was light: it could read articles from the Web saved by “Read Later” services: Read It Later (now Pocket) and Instapaper. There was only one female Amercan English voice. But I made sure that the app was intuitive, smooth, and stable. Twice I threw away what I had built and started over. Granted, I may not have needed to do that if it weren’t for the fact that I had not programmed in 18 years, and I had to learn a new programming language.
What really surprised me was that people actually contacted me. A math teacher uses my app to read out math exams for a student with reading disabilities. A retired lady uses it to listen to blogs about home decoration while she’s doing chores around the house. A writer used it to proof a novel he wrote by listening to it. A college student with dyslexia uses it to listen to reading material from class. A pharmaceutical sales rep uses it to listen to medical research in car, reading that he normally doesn’t have time for. Another gentleman likes to listen to old, non-copyrighted books on obscure topics in ancient history. They’re appreciative for what a $2 app can do for them. And many people wrote to request new features, like additional languages and voices, translation, text highlighting, and ebook reading. I was ready to move on to some other activities that I planned for my year-off, but instead, I dove right back into the thick of it.
After another two months, I now think I have a good piece of software out there. I can say unabashedly that it is the best text-to-speech app in the App Store. The latest release incorporated many customer suggestions. Now 18 languages and 52 voices are available for download. Of course, there are two Norwegian voices. In addition to reading Web pages saved in Read It Later and Instapaper, it can now read content in PDF files, and plain text, ePub, MS Word, MS PowerPoint, Apple Pages, Apple Keynote, RTF files. With Dropbox integration, users can throw what they want to read in Dropbox and open the files directly from my app. It can translate text from any language.
The most time-consuming, and the most fun feature I built is text highlighting. The app can highlight the exact word being spoken, which is great for students with reading disability. And, it can load and read ebooks. I made sure the app can easily load up the King James Bible and read it from start to finish. 72 hours at normal reading speed, by the way.
What next? I don’t know yet. I doubt the app will ever reach a point where it can support the family. So when we go back to Boston in 3 months, I’ll have to find a real job. No matter what happens, though, it’s a piece of work that I’m proud of. I built something that people find useful and I had fun building it. And I didn’t do it to get rich. (I’m charging $2 for the app only because it’s the minimum required by the developer of the text-to-speech engine.) Isn’t that what a sabbatical should be about?
All the thanks go to my wife, Kristin. Thanks for being supportive night after night, enduring the voice of Heather, the default English voice. And for translating the user interface to Norwegian.
The app is called Voice Dream Reader, an offspring of Arctic Dreams.