Voice Design Prompts That Work: 20 Examples
Updated September 3, 2026
TL;DR: a good voice description names six things: age, gender, accent, tone, pace and mood — in that order, in plain words, with one line of sample text so you judge the result on real speech. Each design costs 2,500 credits and is saved permanently; the free tier's sign-up credits cover your first. Twenty examples below, grouped by job.
The structure
[age] [gender] [accent], [tone], [pace], [mood]. Sample: "…" — for example: "Mid-40s woman, neutral American accent, warm and slightly husky, unhurried, reassuring. Sample: 'Let's take this one step at a time.'" Avoid comparisons to real people; describe the voice, not the celebrity.
Narrators (audiobooks, documentary)
- Late-50s man, British RP, dry and precise, measured pace, quietly authoritative.
- Early-40s woman, neutral American, warm low register, steady, intimate — bedtime-story calm.
- 30s man, soft Irish accent, lyrical, unhurried, wistful; for literary fiction.
- 50s woman, Standard Mandarin, composed and weighty, slow, serious; for brand and historical narration.
- 40s man, neutral American, gravelly, slow, grave — trailer and true-crime weight.
Explainers and education
- 30s woman, neutral American, bright and clear, brisk, encouraging; for tutorials.
- 40s man, neutral British, crisp and even, moderate pace, patient; for technical courses.
- 20s woman, Standard Mandarin, clear and friendly, moderate pace, upbeat; for language lessons.
- 30s man, neutral American, conversational and slightly nasal, quick, curious; for tech explainers.
Ads and brand
- 30s man, neutral American, smooth and magnetic, medium pace, confident; premium brand spots.
- 20s woman, light Australian accent, sunny and energetic, fast, playful; DTC product ads.
- 40s woman, neutral American, low and assured, slow, trustworthy; finance and insurance.
- 30s androgynous voice, neutral accent, cool and minimal, moderate, detached; tech launches.
Characters
- Teen boy, neutral American, cracking slightly, quick, nervous and eager.
- Elderly woman, soft Southern U.S. accent, thin and kind, slow, amused.
- Middle-aged man, gruff, heavy, slow, suspicious — the innkeeper.
- Young woman, bright, fast, sarcastic — the sidekick.
- Child (about eight), neutral, high and clear, quick, delighted.
Bilingual and Chinese
- 30s woman, Standard Mandarin with natural English code-switching, smooth, moderate pace, warm; for bilingual channels.
- 40s man, Mandarin with light Taiwanese colour, gentle and precise, unhurried, reassuring; for wellness and meditation.
Iterating
Generate, listen to the sample line, change one attribute at a time — tone before pace, pace before mood — and generate again. Save the candidate you'd use tomorrow, not the one that impressed you today. Designed voices then work everywhere: Text to Speech, character casting in Studio, your ad reads.
Frequently asked questions
How much does a designed voice cost?
2,500 credits per created voice, saved permanently to My Voices. Generating speech with it afterwards uses credits at the normal per-character rate.
Can I describe a famous person's voice?
No. Voice Design is for original voices; describe qualities, not people. Reproducing a real voice requires Voice Cloning with that person's consent.
Can I edit a designed voice later?
Not in place — refine the description and generate a new candidate, then save that one.
Which languages can a designed voice speak?
It generates through the multilingual Ultra engine, so English, Chinese and the other supported languages are available; specify the accent you want.