One morning in early April 2018, I woke up in the pre-dawn hours and couldn’t get back to sleep. I was thinking about voice technology and specifically about content authors formatting text with Speech Synthesis Markup Language (SSML). With the thoughts swimming around in my head, I just couldn’t get back to sleep. I needed to get these ideas “on paper” so to speak.

Months earlier I had created a website, SSML Guru, that documented various differences in how voice platforms implemented (or didn’t) the full SSML specification. Having implemented various voice applications, I found myself constantly referring to documentation to remember the tags and attributes of SSML. I was thinking about frameworks that would allow coding of a voice app once and then being able to deploy that as an Alexa Skill or Action on Google Assistant. What about the “no-code” visual designers not targeted at developers at all?

Since 2015, developers have coded text-to-speech responses for modern voice assistant platforms (Alexa, Google Assistant, Cortana, etc.) These responses were typically coded into the project or stored in a database. Over time, more designers have become involved and the only available way to format responses remains SSML. I was thinking about the next steps with content moving to Content Management Systems (CMS) and content authors being more prevalent. Was SSML the right tool for them? I didn’t think so.

I thought about the complexities of the XML-based syntax of SSML and how it was not unlike HTML. As an author writing content for publication on WordPress, I wouldn’t write HTML. I would write text. And when I needed to format that text, I would use… Markdown! Content authors use markdown! What a simple concept. Why hadn’t I thought of that sooner?

What content authors creating for this new voice age needed was a markdown of their own. Speech Markdown!

That was the core thought that got me so excited that spring morning in 2018. I had to write it down. I needed to experiment with different syntax options to see what would make sense. For hours I worked through these ideas and documented as much as I could.

That was 14 months ago. Over these months, I have changed much of the syntax that I had originally documented. Some additions have been made for platform-specific SSML (such as voice for Amazon Alexa) so new markdown was created. Even though work, life, and other challenges got in the way, I have always believed in Speech Markdown. Any doubts have been about me being up to the task. After putting in hundreds of hours into thinking about, experimenting, and documenting, will others see the value in Speech Markdown? There is only one way to find out.

Those many months ago, I thought Speech Markdown would be useful. That its time had come. I believe that even more so today. From Alexa Blueprints and visual tools to better coding frameworks, the time has come for content authors, designers, and developers to have a markdown of their own.

To the amazing VoiceFirst community out there and the many friends that I have made. Thank you!

Speech Markdown is for you! It is open source. Come join the project and let’s work together to advance voice technology.


Mark Tucker @marktucker