How to create a designed voice
AI Voice Designer is useful when you know the voice style you want but do not need to clone a real person. By the end of this guide, you will have a designed voice saved in My Voices and ready to test with real script copy.
Before you start
- Confirm whether Personal or a Workspace is selected before generating. The completed voice belongs to that selected space; a Workspace voice is shared with its members.
- Know the job the voice needs to do: narrator, host, character, support agent, teacher, or presenter.
- Write down the voice traits before opening the tool: tone, pace, accent, age range, energy, and delivery style.
- Use Voice Clone instead only when you have clean source audio and documented permission for a real voice.
Create it in Giggy
- In the left sidebar, open AI Voice Designer.
- Enter a clear voice name.
- Describe the voice with literal traits: tone, pace, accent, age range, delivery, and use case.
- Choose the language.
- Select Generate.
- Review the generated voice in My Voices before using it in Text to Speech.
Improve the result
- Describe vocal traits directly: pace, tone, age range, accent, energy, and delivery.
- Name the intended use case in the prompt when the delivery matters.
- Avoid vague prompts like "make it good"; explain what the voice should do.
- Do not request a specific real person unless you have the required rights.
- Test the designed voice with real project copy before using it at scale.
- After the voice is created, test expression tags in Text to Speech instead of trying to put non-verbal tags into the voice description.
- Generate a few short previews and choose the voice that handles your actual script most clearly.
Examples
- Good description: "Warm documentary narrator, adult voice, calm pace, clear American accent, confident but not dramatic, suited for product explainers."
- Character direction: "Bright animated sidekick, young adult energy, quick pace, playful delivery, clear consonants for short dialogue."
- Brand direction: "Polished luxury host, measured pace, low energy, subtle warmth, suited for premium product videos."
Common mistakes
- Writing vague prompts such as "make it good" or "professional" without vocal traits.
- Trying to imitate a specific real person instead of describing an original voice style.
- Judging the voice from a generic sample instead of testing it with your actual script.
Use cases
Brand and series voices
- Create a reusable voice style for product videos, recurring narration, courses, or branded social formats.
- Create a fictional presenter when a public library voice is too generic.
Character and concept design
- Design character voices for scripts, stories, games, prototypes, and learning content.
- Test different voice personas against the same script before choosing one direction.
No-source-audio workflows
- Make a custom voice style without uploading source audio.
- Create voice direction before deciding whether an authorized clone is actually necessary.