Voice Design Prompts for Creators
Copy-paste prompt patterns for creating role-based AI voices for narration, product demos, courses, and character work.
Voice Design in Murmur is about describing the kind of voice you need for a project. You write a role and style prompt, preview the result, and save useful voices into your workflow. It is not a tool for copying a celebrity, imitating a private person, or pretending a voice belongs to someone else.
Good voice prompts describe function, tone, audience, pacing, and context. A prompt like "sound like a famous actor" is the wrong direction. A prompt like "calm documentary narrator for a science explainer, measured pace, warm but not dramatic" gives the model practical information.
The Prompt Structure That Works
- Role: tutorial host, product narrator, field reporter, course instructor, or audiobook guide.
- Tone: calm, direct, warm, precise, energetic, conversational, serious, or friendly.
- Pace: slow, measured, brisk, punchy, relaxed, or steady.
- Use case: YouTube explainer, onboarding video, course lesson, ad read, podcast intro, or audiobook chapter.
- Boundary: describe a useful role instead of a real person or celebrity.
Start with one clear sentence, generate a preview, then change one variable at a time. If the voice is too flat, add a tone cue. If it rushes, ask for a more measured pace. If it feels too theatrical, ask for restrained delivery.
Prompt Examples
- Calm tutorial host for a software walkthrough, clear pronunciation, steady pace, friendly but focused.
- Warm product narrator for a launch video, confident delivery, medium pace, polished but not salesy.
- Documentary-style explainer voice for a science channel, thoughtful tone, measured pacing, minimal drama.
- Short-form ad voice for a creator tool, energetic and direct, quick pace, clear call to action.
- Course instructor for beginner lessons, patient tone, slower pace, precise pronunciation, low hype.
- Podcast intro voice for an independent creator show, welcoming tone, relaxed pace, crisp ending.
- Field narrator for a travel video, observant and natural, lightly curious tone, conversational pace.
- Audiobook narrator for nonfiction chapters, even pacing, mature tone, clear paragraph transitions.
How to Refine a Prompt
When a generated voice is close but not right, avoid rewriting the whole prompt. Change one part. If the role is correct but the tone is wrong, keep the role and adjust the tone. If the tone works but the delivery drags, change only the pace.
- Write a one-sentence prompt using role, tone, pace, and use case.
- Preview it with a short line from the real project.
- Write down what feels wrong in plain language.
- Change one prompt variable and preview again.
- Save the voice only when it works on the real script, not just a demo sentence.
Weak vs Better Prompts
| Weak prompt | Better prompt | Why |
|---|---|---|
| Sound like a movie trailer voice | Dramatic trailer narrator for a 15-second product teaser, deep tone, slow build | It describes style without naming a person |
| Make it viral | Energetic short-form ad voice, quick pace, direct delivery, clear final line | It explains the job |
| Professional voice | Confident product narrator for a B2B onboarding video, steady pace, neutral warmth | It defines audience and context |
| YouTube voice | Conversational explainer host for a faceless YouTube channel, curious tone, medium pace | It narrows the format |
Build a Small Voice Library
Creators usually need a few dependable voices, not dozens of random experiments. A simple library might include one tutorial host, one product narrator, one short-form ad voice, one calm explainer, and one long-form narrator. Save voices that survive real scripts. Ignore voices that only sound impressive on tiny samples.
Murmur costs $49 one-time and does not use a cloud credit meter, which makes iteration easier. There is no free trial, so it is best evaluated as a paid local Mac tool for creators who expect to generate voiceovers regularly.
Design reusable voices for creator work.
Use Murmur to create role and style based voices for narration, tutorials, product videos, and long-form projects on your Mac.
macOS 14+ · Apple Silicon required · 7-day refund policy