What Is SSML (Speech Synthesis Markup Language)?

SSML is a W3C standard. Version 1.0 became a W3C Recommendation in 2004 and version 1.1 in 2010. Major cloud text-to-speech services accept SSML, though each supports its own subset of the tags and adds a few of its own.

• break adds a pause, for example a one-second silence.
• prosody changes speaking rate, pitch or volume.
• say-as says how to read a value, such as a date, a number or a spelled-out word.
• phoneme gives an exact pronunciation using a phonetic alphabet.
• sub substitutes one written form for another, such as expanding an abbreviation.
• emphasis stresses a word.

Writing raw XML tags inside a script is fiddly, easy to break with a missing quote or bracket, and different providers treat the same tags differently. Many people only ever need pauses and the occasional pronunciation fix.

You do not have to write SSML. Pauses are placed with an Insert pause button, which adds a break tag (for example a one-second pause) that you can also type by hand, from 0.1 seconds up to 10 minutes. Pronunciation fixes are done with pronunciation rules rather than markup, and speed is a slider from 0.5× to 2×. GetMeAudio does not aim to expose every SSML tag.