TTS Best Practices

Text formatting

Do this	Instead of
"Hello, how are you?"	"Hello how are you"
"In 2024, revenue grew."	"In 2024 revenue grew"
"Visit example dot com"	"Visit example.com"
"three hundred dollars"	"$300" (if voice mispronounces)

Pacing and pauses

Use commas for short pauses
Use periods for full stops
Use em dashes (—) for dramatic pauses
Add an extra line break between paragraphs for breathing room

For a pause of an exact length, use a <break /> tag inline:

Let me think <break time="1.5s" /> Yes, I've thought about it.

Seconds or milliseconds both work (1.5s = 1500ms), up to 20 per request, in any language. Break tags never show up in captions or SRT. See Voice Steering for the full list of inline controls.

Emphasis

ALL CAPS adds emphasis: "This is IMPORTANT" → stressed word
Exclamation marks add energy: "Let's go!"
Avoid ALL CAPS entire sentences — sounds robotic

Pronunciation fixes

If a word is mispronounced, try:

Spelling it phonetically: "GIF" → "jif"
Breaking it up: "API" → "A P I"
Writing it out: "Dr." → "Doctor"

Voice selection

Match voice tone to content — don't use an energetic voice for meditation content
Professional voices are best for monetized/commercial content
Check the voice sample before generating a long script

Long-form content

Split into paragraphs of ~300 characters each
Use the Audiobook editor in Studio for easy line-by-line editing — see How to Make an Audiobook
Review each chapter before exporting the full project