Grok TTS Speech Tags: All 26 Tags With Examples
Every speech tag Grok text to speech understands: 14 inline tags like [pause] and [laugh], 12 wrapping tags like <whisper>, each with an example to try.
Last updated:
Grok text to speech understands 26 speech tags: 14 inline tags written as [tag], which add a sound or a pause at that spot, and 12 wrapping tags written as <tag>…</tag>, which change how the wrapped phrase is spoken. xAI sorts them into seven groups: pauses, laughter and crying, mouth sounds, breathing, volume and intensity, pitch and speed, and vocal style. The list below matches xAI's text to speech docs as checked on 17 October 2026. Paste any example into the free Grok text to speech tool to hear it in any of the 28 voices, with no sign-up and no API key.
How the two kinds of tags work
- Inline tags go where the sound would happen in a real conversation:
Really? [laugh] That's incredible! - Wrapping tags need an opening and a matching closing tag, and work best around a complete phrase:
<whisper>It is a secret.</whisper> - Tags combine. Wrapping tags nest (
<slow><soft>Goodnight, sleep well.</soft></slow>), and inline and wrapping tags can share a sentence. - Tags are not spoken, but they count. They are part of the text you send, so they count toward the character limit (1,000 per clip here) and toward what xAI bills.
The example sentences in the tables are ours; the tag names and groups are xAI's.
Pauses
Timing tags. Use [pause] or [long-pause] to let a thought land or to add dramatic timing. xAI lists [hum-tune] in this group too.
| Tag | Kind | Example |
|---|---|---|
[pause] | Inline | Let me check that for you. [pause] Okay, your order shipped this morning. |
[long-pause] | Inline | And the winner is… [long-pause] the blue team! |
[hum-tune] | Inline | [hum-tune] Sorry, that song has been stuck in my head all day. |
Laughter and crying
| Tag | Kind | Example |
|---|---|---|
[laugh] | Inline | Really? [laugh] That's the best excuse I've heard all week. |
[chuckle] | Inline | He said he'd be five minutes. [chuckle] That was an hour ago. |
[giggle] | Inline | Stop it, you're making me blush. [giggle] |
[cry] | Inline | I just miss her so much. [cry] Sorry, give me a second. |
Mouth sounds
| Tag | Kind | Example |
|---|---|---|
[tsk] | Inline | [tsk] You forgot to save the file again. |
[tongue-click] | Inline | [tongue-click] Right, let's try that one more time. |
[lip-smack] | Inline | [lip-smack] Mm, that soup was worth the wait. |
Breathing
| Tag | Kind | Example |
|---|---|---|
[breath] | Inline | Okay. [breath] Here's what we're going to do. |
[inhale] | Inline | [inhale] Smell that? Fresh bread, straight from the oven. |
[exhale] | Inline | We made it to the top. [exhale] What a climb. |
[sigh] | Inline | [sigh] Monday again already. |
Volume and intensity
Wrapping tags. <build-intensity> and <decrease-intensity> work best around a whole sentence, so the change has room to happen.
| Tag | Kind | Example |
|---|---|---|
<soft> | Wrapping | <soft>It's okay. You're safe now.</soft> |
<whisper> | Wrapping | Come closer. <whisper>I know where the key is hidden.</whisper> |
<loud> | Wrapping | <loud>Everybody, please take your seats!</loud> |
<build-intensity> | Wrapping | <build-intensity>The crowd rises, the clock ticks down, and he shoots!</build-intensity> |
<decrease-intensity> | Wrapping | <decrease-intensity>The music fades, the lights go down, and the room falls quiet.</decrease-intensity> |
Pitch and speed
Wrapping tags. They change only the wrapped phrase; for the pace of the whole clip, use the speed setting (0.7× to 1.5×) instead.
| Tag | Kind | Example |
|---|---|---|
<higher-pitch> | Wrapping | <higher-pitch>Oh my gosh, is that a puppy?</higher-pitch> |
<lower-pitch> | Wrapping | <lower-pitch>Nobody leaves this room until we find out who did it.</lower-pitch> |
<slow> | Wrapping | <slow>Breathe in for four seconds, and out for four.</slow> |
<fast> | Wrapping | <fast>Offer ends Sunday, terms apply, see store for details.</fast> |
Vocal style
| Tag | Kind | Example |
|---|---|---|
<sing-song> | Wrapping | <sing-song>Somebody's got a birthday today!</sing-song> |
<singing> | Wrapping | <singing>Good morning, sunshine, it's time to rise.</singing> |
<emphasis> | Wrapping | Before you leave, <emphasis>check that the oven is off.</emphasis> |
Tips for natural results
These follow xAI's own advice in its docs:
- Pair tags with punctuation.
Really? [laugh] That's incredible!sounds more natural than several tags in a row. - Wrap phrases, not single words.
<whisper>It is a secret.</whisper>reads better than whispering one word. - Nest for combined effects.
<slow><soft>Goodnight, sleep well.</soft></slow>is slower and softer. - Use pauses for timing.
[pause]and[long-pause]give a punchline or a reveal room to land.
Try them without writing code
The Grok text to speech tool has buttons above the text box that insert [pause], [laugh], [sigh], <whisper> and <emphasis> at the cursor (a wrapping tag wraps whatever text you have selected), and a menu with the other 21 tags. The samples there (storytelling, bedtime story, comedy bit) show several tags working together. Every voice accepts the same tags, so you can compare how Ara, Eve or any of the 28 voices deliver the same line.
Using tags through the API
Tags go straight into the text field of POST https://api.x.ai/v1/tts; there is no separate parameter for them. The request body, voice ids, output formats and code samples are in the Grok Voice API guide, and the pricing page shows what tagged text costs per million characters.