Tell your AI voice exactly how to sound.
Levizr's system prompt feature lets you define the full personality, tone, pacing, and style of your AI voice — in plain English. No audio editing. No manual post-processing. Just describe the voice you want, and the neural model delivers it.
What the system prompt controls
Unlike a speed slider or a pitch dial, a system prompt is a holistic instruction. It shapes the entire character of the voice across the generation — not just a single acoustic parameter. Here is everything you can influence with a well-written prompt.
Speaking Speed & Pacing
Instruct the voice to speak fast and snappy for ads, or slow and deliberate for educational content. System prompts give you nuanced control over rhythm that speed sliders alone cannot replicate.
Tone & Formality Level
Shift between ultra-casual, conversational, professional, and formal — just by describing who the speaker is in the prompt. 'Speak like a doctor explaining a diagnosis' produces very different output from 'Speak like a podcast host chatting with a friend.'
Personality & Character
Define a specific persona for the voice. 'Friendly startup founder', 'wise elder storyteller', 'enthusiastic fitness coach' — the neural model adapts its delivery to the persona you describe.
Emphasis & Stress Patterns
Tell the model which types of words to stress. 'Place emphasis on action words and numbers' or 'slow down on product names' gives you director-level control without any manual audio editing.
Audience Awareness
Describe the listener. 'This is for a 10-year-old audience' or 'This is for C-suite executives' shapes vocabulary choices, pacing, and tonal weight in the generated audio.
Emotional Baseline
Set a default emotional state for the entire script. 'Maintain a warm and hopeful tone throughout' acts as a floor — any emotion tags in your script then add variation on top of this baseline.
System prompt examples — ready to use
Copy any of these directly into the Levizr Studio system prompt field and hear the difference in your generated audio. Each prompt is designed for a specific content format.
Prompt
"Speak like an enthusiastic YouTube educator — fast-paced, clear, and engaging. Keep the energy high throughout."
What changes in the audio
Voice becomes noticeably faster, with rising intonation on key points and natural emphasis on action words.
Prompt
"Professional and authoritative tone. Speak clearly and confidently. Avoid sounding casual or overly enthusiastic."
What changes in the audio
Voice adopts measured pacing, consistent volume, and a formal delivery suited for boardroom-style audio.
Prompt
"Warm, playful storyteller voice. Sound like you're reading to a 6-year-old — gentle, expressive, and fun."
What changes in the audio
Voice softens and becomes more melodic, with exaggerated rises and falls that hold a young listener's attention.
Prompt
"Dark, slow, and unsettling. Speak like a horror movie narrator — pause dramatically and let the silence do the work."
What changes in the audio
Voice slows down significantly, pitch drops, and the delivery feels heavy and suspenseful.
Prompt
"Calm, slow, and deeply soothing. Breathe naturally between sentences. Sound like a mindfulness instructor."
What changes in the audio
Voice becomes noticeably slower with gentle pausing between phrases — ideal for relaxation or sleep content.
Prompt
"Powerful, passionate, and inspiring. Speak like a high-performance coach pushing someone toward their goals."
What changes in the audio
Voice gains volume and intensity, with strong emphasis on action verbs and an energizing forward momentum.
Don't want to write a prompt? Use the AI persona builder.
The Advanced Configuration modal in Levizr Studio lets you select structured settings — age group, mood, regional accent, and speaking speed — and then automatically generates a detailed, rich system prompt based on your choices. The AI persona builder sends your selections to an internal API that returns a fully written prompt you can use immediately or edit to your preference.
This is the recommended starting point if you are new to prompt-based voice control. The auto-generated prompts are verbose and detailed — much more effective than the short one-liners most users write manually.
Open Advanced Configuration →Shapes vocabulary pacing and energy to match the speaker's perceived age.
Sets the emotional starting point for the generated voice persona.
Fine-tune playback rate from half-speed narration to double-speed summaries.
PRO — Applies a regional accent overlay to compatible voices. Ideal for localized content.
System prompt adds 5 credits to your generation
When a system prompt is active, Levizr adds a flat 5 credits on top of your per-character script cost. If the system prompt field is empty, those 5 credits are not charged. The Advanced Configuration persona builder uses the same 5-credit system prompt cost — the AI generation of the prompt itself is free.
Frequently Asked Questions
What is a system prompt in Levizr TTS?
A system prompt is an instruction you give to the neural model before it reads your script. It defines the voice's personality, speaking style, tone, pace, and emotional baseline — without you needing to edit the actual audio. Think of it as a brief you give to a voice actor before they record.
Does the system prompt cost extra credits?
Yes — using a system prompt adds a flat 5 credits to the generation cost. This is in addition to the per-character cost of your script text. If your system prompt is empty, those 5 credits are not charged.
What is Advanced Configuration and how is it different from the system prompt?
The Advanced Configuration panel lets you pick structured settings — age group, mood, accent, and speed — and then auto-generates a rich system prompt based on those choices. It is a guided way to write a system prompt without having to describe everything manually. The AI generates a detailed persona description from your selections.
Can I combine a system prompt with emotion tags?
Yes, and this is the recommended approach. The system prompt sets the overall tonal baseline for the entire generation, while emotion tags apply local variation to specific phrases. For example: system prompt sets a "warm, professional tone" and an [excited] tag adds a burst of energy on a specific line.
Does the system prompt affect all voices equally?
Different voices respond differently to the same prompt. A system prompt asking for "deep and authoritative delivery" will produce a more noticeable change with a voice like Charon (already firm and deep) than with Leda (youthful and light). It is worth previewing different voice and prompt combinations to find the best match.
Is the system prompt saved between sessions?
Yes. Levizr automatically saves your last-used system prompt to localStorage and restores it when you return. You can also pass it via URL using the ?sp= query parameter if you want to share a specific configuration with someone else.