Best PromptsCustom Voice Design
Create an original voice with ElevenLabs Voice Design: the attribute formula that works, preview-text strategy, and one-variable iteration.
Design a custom voice with ElevenLabs Voice Design.
Who the voice is for: [THE USE: "narrator for my history channel" / "the villain in my audio drama" / "my product's onboarding voice" / "ad reads for my podcast"]
The character in 3 words: [e.g. "wise, dry, unhurried" / "bright, quick, trustworthy"]
Build the voice prompt with the attribute formula: specific beats poetic:
"[AUDIO QUALITY]: Perfect studio audio quality. [DEMOGRAPHICS]: [GENDER], [AGE BAND: young adult / in their 40s / elderly]. [ACCENT (the word 'thick' or 'slight' + region: 'slight Irish accent' / 'thick New York accent') or omit for neutral]. [TIMBRE, 2-3 from: deep, warm, gravelly, smooth, buttery, raspy, airy, nasal, bright, low-pitched]. [PACING: 'calm, measured conversational pace' / 'fast-paced and energetic' / 'slow and deliberate with natural pauses']. Persona: [2-5 WORDS: 'retired sea captain telling stories' / 'sharp museum docent']. Emotion: [2-3 ADJECTIVES: 'reflective, gently amused']."
What NOT to put in the prompt (these degrade the voice):
- Audio effects: reverb, echo, telephone, radio, distortion. Those are post-production, not voice traits; in the prompt they just add mud. (Exception: "poor audio quality" as a deliberate character choice, e.g. an old broadcast voice.)
- Vague vibes: "exotic", "unique", "interesting"; the model needs physical attributes.
- "Accent" when you mean emphasis or attitude. It will shift the dialect.
The preview text (matters more than people think, the voice is judged performing THIS): write 2-3 full sentences in the register the voice will actually speak: [PASTE A REAL SAMPLE OF WHAT IT WILL READ: the actual narration opening / the character's actual lines / the actual ad copy]. Energetic copy auditions a calm voice falsely, and vice versa.
Iteration protocol: generate 3 previews per prompt; change ONE attribute at a time between rounds ("same prompt, pacing slower" / "same, timbre warmer"): multi-attribute changes make it impossible to know what moved. When one lands: save it, then control per-generation delivery with the TTS settings (stability, style) and inline audio tags; don't re-roll the voice to fix a single reading.
Use-case starting points:
- Narrator: "Perfect studio quality. [Gender], middle-aged. Warm, smooth, low-pitched. Calm measured pace. Persona: seasoned documentary narrator. Emotion: quietly authoritative, curious."
- Character: lean hard on persona + emotion + one exaggerated texture trait.
- Ad read: "energetic, upbeat, crisp diction, fast-paced, confident" + perfect audio quality.More audio prompts
Build me a style prompt for [SUNO / UDIO / ELEVENLABS MUSIC]. What I want it to sound like: [DESCRIBE IT IN YOUR OWN WORDS, INCLUDING ARTISTS YOU HAVE IN MIND. I will translate those into describable qualities rather than using the names.] Where it will be used: [BACKGROUND FOR A VIDEO / A SONG PEOPLE LISTEN TO / A PODCAST INTRO / A GAME
Suno Style Prompt Builder
Build a precise style field from era, instrumentation, vocals, and production instead of artist names, plus exclusions that fix a bad take.
Audiobeginner
Write a jingle for [BRAND]. What the brand does, in plain words: [ONE SENTENCE] The one thing to remember: [THE NAME, THE OFFER, OR THE PHONE NUMBER] Audience: [WHO] Brand personality: [THREE WORDS] Where it plays: [RADIO / TV / SOCIAL PRE-ROLL / IN STORE / PODCAST READ] Lengths needed: [e.g. 5 seconds, 15 seconds, 30 seconds] Genre and
Brand Jingle
Write a jingle or brand anthem people remember: one hook, a name that lands on the strong beat, and versions cut for every ad length.
Audiointermediate
Direct the narration for this text. Text to narrate: """ [PASTE THE CHAPTER OR SECTION] """ Genre: [LITERARY FICTION / THRILLER / MEMOIR / BUSINESS NONFICTION / CHILDREN'S] Narrator: [FIRST PERSON AS THE PROTAGONIST / THIRD PERSON OMNISCIENT / THE AUTHOR READING THEIR OWN WORK] Listener and where they listen: [e.g. "commuters, in traffi
Audiobook Narration
Direct long-form narration that stays consistent over hours: pace, character voices, pronunciation list, and chapter-level consistency notes.
Audioadvanced