Every documented v3 audio tag, grouped by what it does — and a dialogue editor right here to hear each one. Click a tag to copy it, paste it into a line, and generate.
Set how the line feels.
Change volume, speed and intensity.
Non-verbal sounds a person makes.
Shift the voice's character without switching voices.
Make multi-speaker dialogue feel real.
Short effects inside the audio.
Free credits on sign-up · no card needed · Turbo costs 3 credits per 1,000 characters
Tags sit inline in your script, in square brackets, and direct how the next words are performed. You can stack them — “[excited] [shouts] We won!” — and change them mid-line. They're directions, not guarantees: the result depends on the voice, so a naturally deep, calm narrator will interpret [shouts] differently from an energetic announcer.
On this site, tags work in Dialogue mode, which runs Dialogue v3 at 7 credits per 1,000 characters. Tags count toward the characters.
Put the tag in square brackets right before the words it should affect: "[whispers] Don't turn around." Tags are case-insensitive; lowercase is the convention.
v3 reads a tag as text when the voice can't perform it — for example asking a soft, calm voice to [shouts] repeatedly. Pick a voice with a wider emotional range, or use a gentler tag.
No. Audio tags are a v3 feature. On this site they work in Dialogue mode, which runs v3.
It lists the tags ElevenLabs documents. v3 understands many other natural-language directions too — try your own, like [nervously] or [trying not to laugh].
No. We're an independent service that runs ElevenLabs models among others. ElevenLabs is a trademark of its owner.
Tag list based on ElevenLabs' published documentation. Not affiliated with or endorsed by ElevenLabs.