Short answer: to get whispered text to speech, put your text inside a <whisper>…</whisper> tag and generate it with Grok text to speech. The box above already does this: Whisper all text wraps everything you type, so you can paste a line, pick one of 28 voices and press Generate. Turn the button off to whisper only part of a script. Trying it needs no account.
How to make a whisper voice
- Paste or type your text in the box. The sample line is already whispered; replace it with your own.
- Keep Whisper all text on to whisper everything. The box shows the tag around your text, so you always see what gets sent.
- Pick a voice. All 28 voices accept the whisper tag; the clips above use Eve.
- Press Generate and listen. Signed in, you can download the result as an MP3.
To whisper just one phrase, turn Whisper all text off, select the phrase and press the <whisper> chip under Add expression. That wraps only your selection:
I opened the door, [pause] and froze. <whisper>Something was wrong.</whisper>
What the whisper tag does
<whisper> is one of 26 speech tags in xAI's text to speech docs. It is a wrapping tag: it changes how the words between <whisper> and </whisper> are spoken, and the tags themselves are not read out. In our two pre-recorded clips of the same line in the same voice, the whispered one came out about 8 dB quieter on average. Expect the same voice speaking quietly, not a different voice.
Ideas to try
These are suggestions, not tests we ran for each use:
- Bedtime stories: combine
<whisper>with<soft>and<slow>for the last line of a story. - Horror and suspense narration: keep the narration normal and whisper the line that should land, with a
[pause]before it. - Secrets in dialogue: whisper one character's aside while the other speaks normally.
- Quiet voiceovers: whisper a short intro or outro, then switch back to a normal read.
Tips for a better whisper
- Whisper whole sentences. A wrapped phrase reads more naturally than a single whispered word.
- Use punctuation for rhythm. Commas and full stops give the whisper places to breathe;
[breath]and[pause]add more. - Try two or three voices. The same text can sound quite different from voice to voice; the voices page has every voice reading one sentence.
- Mix in other tags.
[sigh],[chuckle]and<slow>work in the same text. The full list is in the speech tags guide, and the text to speech with emotion page combines them into one-click moods. The shouting text to speech page is this page's opposite, with<loud>.
Limits, price and privacy
Signed out, you get 3 previews a day of up to 200 characters, played on the page; a free account adds MP3 download, and a Google account raises the limit to 1,000 characters per clip. The tag characters count toward the limit. Your text goes to xAI's model through our Cloudflare function to make the audio and is not stored.
If you build your own app, the same model costs $15 per million characters through xAI's API; see the pricing calculator. There is no separate whisper price; xAI bills text to speech per character.
Last updated: