Skip to content
English - Australia
  • There are no suggestions because the search field is empty.

Otto Custom Voices - Prompting Guide

Otto Custom Voices are useful when the standard Otto voices do not quite match your venue, brand, customer base, or preferred accent.

Prompt Structure

Voice creation is iterative. A more specific prompt usually gives Otto more useful direction, especially when you care about accent, age, tone, pacing, or delivery style. A short prompt can still work well if you want a simple, neutral voice.

For consistent results, structure your prompt like this:

  • Native [language and dialect]. [Gender or voice style], [age range]. [audio quality].
  • Persona: [2 to 5 words]. Emotion: [2 to 3 adjectives].
  • [1 to 2 sentences about tone, pitch, pacing, and delivery.]

General example:

Native Australian English only. Female, 30-40. Excellent quality. Persona: friendly restaurant host. Emotion: warm, attentive, confident. Smooth, natural voice with a mid pitch, clear pronunciation, relaxed conversational pacing, and gentle emphasis on order details. Sounds professional and local without sounding overly formal or theatrical.

Decide how detailed your prompt should be

Use more detail when you want a specific result, such as a particular accent, venue personality, age range, or speaking pace.

Use a shorter prompt when you want a broadly usable voice. For example:

Calm male restaurant host with a natural Australian accent, excellent audio quality, and relaxed conversational pacing.

Avoid adding details that clash with each other. For example, a prompt that asks for a calm, gentle voice and a loud, high-energy delivery may produce inconsistent previews.

Language and dialect

Put the language and dialect first. This helps Otto stay anchored to the right regional sound.

Useful examples:

  • Native Australian English.
  • Native New Zealand English.
  • Native British English.
  • Native Spanish, European Spanish.
  • Native Arabic with a soft Gulf accent influence.

If the regional version matters, say so clearly. For example, use "Australian English, not American English" or "European Spanish, not Latin American Spanish".

Audio quality

For restaurant calls, clear audio is usually the safest choice. Include a quality phrase in your prompt if you want the voice to sound polished and easy to understand.

Useful quality phrases:

  • Good quality.
  • Very good quality.
  • Excellent quality.
  • Studio quality.
  • Broadcast quality.
  • Perfect audio quality.

For most Otto agents, use "excellent quality", "studio quality", or "broadcast quality".

Avoid special effect words such as "reverb", "echo", "phone", or "tape". These can make the voice sound less clean. Also avoid asking for low-quality audio unless you have a very specific reason. Phrases such as "low-fidelity audio", "poor audio quality", "sounds like a voicemail", or "muffled and distant" are usually not suitable for a customer-facing restaurant agent.

Age

Age helps define the maturity, texture, and energy of the voice. Be specific enough to guide the result.

Useful age phrases:

  • Young adult.
  • In their 20s.
  • Early 30s.
  • 30-40.
  • Middle-aged.
  • In their 40s.
  • Older voice.
  • Elderly voice.

Gender or voice style

You can describe the voice by gender, or you can describe the sound without relying only on gender.

Useful examples:

  • Male.
  • Female.
  • Gender neutral.
  • Lower-pitched female voice.
  • Deep, resonant male voice.
  • Soft, mid-pitched neutral voice.

If the sound matters more than the identity, describe the pitch, weight, and texture directly.

Tone, timbre, and pitch

Tone and timbre describe the physical quality of the voice. They are different from emotion.

Useful tone, timbre, and pitch phrases:

  • Deep.
  • Low-pitched.
  • High-pitched.
  • Smooth.
  • Rich.
  • Warm.
  • Mellow.
  • Gravelly.
  • Raspy.
  • Airy.
  • Breathy.
  • Resonant.
  • Light.
  • Thin.
  • Nasally.
  • Shrill.
  • Tinny.
  • Metallic.

For restaurant agents, useful combinations often include "warm and smooth", "clear and mid-pitched", "calm and resonant", or "light and friendly".

Persona or role

Persona helps Otto understand the kind of person the voice should feel like.

Useful restaurant personas:

  • Friendly restaurant host.
  • Helpful takeaway assistant.
  • Calm reservations coordinator.
  • Polished fine-dining host.
  • Warm family restaurant host.
  • Efficient front-of-house team member.
  • Confident customer service agent.

Keep the persona short. Two to five words is usually enough.

Emotion

Emotion describes the attitude the voice should carry.

Useful emotion words:

  • Warm.
  • Friendly.
  • Reassuring.
  • Attentive.
  • Confident.
  • Polite.
  • Calm.
  • Energetic.
  • Helpful.
  • Professional.
  • Dry.
  • Sarcastic.

For most Otto agents, use emotions that fit customer service calls, such as warm, attentive, confident, polite, and helpful.

Accent

Accent affects how local, familiar, or distinctive the voice sounds. Be clear and deliberate if accent matters for your venue.

Use "accent" when you mean a regional way of speaking. If you mean the rise and fall of the voice, use "intonation". If you mean which words are stressed, use "emphasis" or "delivery".

Useful accent phrases:

  • Slight Australian accent.
  • Neutral Australian accent.
  • Regional Australian accent.
  • Broad Australian accent.
  • Thick French accent.
  • Soft Irish lilt.
  • Crisp British accent.
  • Neutral American accent.

When you want a noticeable accent, words such as "thick" or "broad" may work better than "strong". When you want a lighter accent, use "slight", "soft", "neutral", or "subtle".

Avoid vague words such as "foreign" or "exotic". They are not precise enough and can create inconsistent results.

Combine accent with other traits for better control. For example:

Middle-aged male voice with a neutral Australian accent, warm timbre, and relaxed conversational pacing.

Getting the Australian accent right

If you want an Australian voice, make Australian English the first instruction in the prompt.

Best starting phrases:

  • Native Australian English only.
  • Native Australian English speaker.
  • Australian English, not American English.
  • Natural Australian accent.
  • Neutral Australian accent.
  • Thick Australian accent.
  • Local Australian pronunciation and intonation.
  • Strong Aussie pronunciation.
  • Relaxed Australian delivery, without heavy slang.
  • Non-rhotic R sounds.
  • Avoid all American pronunciation.

For a clearly Australian-sounding voice, this structure works well:

  • [Accent strength]. [Gender] native Australian English speaker. [Age range]. [Audio quality].
  • [Persona]. [Emotion words].
  • [Australian pronunciation, rhythm, R sounds, and delivery]. [What to avoid].

Recommended Australian prompt:

Thick Australian accent. Female native Australian English speaker. 20-30. Excellent quality.

Bubbly, energetic local restaurant host. Friendly, confident, helpful. Strong Aussie pronunciation, relaxed Australian rhythm, non-rhotic R sounds, and clear conversational delivery. Avoid all American pronunciation.

In this prompt, "non-rhotic R sounds" helps steer the voice away from American-style R pronunciation. It is a useful phrase when an Australian preview keeps sounding American.

Choose how strong the accent should be:

  • Use "subtle Australian accent" if you want the accent to be present but polished.
  • Use "neutral Australian accent" if you want a local sound that is softer and more polished.
  • Use "natural Australian accent" if you want the voice to sound local without sounding exaggerated.
  • Use "thick Australian accent" if the preview needs to sound more clearly Australian.
  • Use "regional Australian accent" if you want a more location-specific feel.
  • Use "broad Australian accent" or "thick Australian accent" only if you want the accent to be very noticeable.

Avoid writing only "Aussie voice". It can be too vague. Use more precise wording such as "native Australian English speaker", "strong Aussie pronunciation", "relaxed Australian rhythm", and "avoid all American pronunciation". Also avoid relying on slang to create the accent. A restaurant agent should usually sound local, clear, and natural rather than like a character.

If the preview sounds American:

  • Move "native Australian English speaker" to the first sentence.
  • Add "avoid all American pronunciation".
  • Add "Australian pronunciation and intonation".
  • Add "relaxed Australian rhythm".
  • Add "non-rhotic R sounds".
  • Use "thick Australian accent" if softer phrases are not strong enough.

If the preview sounds British:

  • Add "not British English".
  • Add "local Australian pronunciation".
  • Avoid phrases such as "crisp", "proper", or "formal" if they are pulling the voice toward a British sound.

If the accent is too broad:

  • Replace "thick" or "broad" with "neutral", "natural", or "subtle".
  • Add "not exaggerated".
  • Add "without heavy slang".

If the accent is right but the delivery feels flat:

  • Add "gentle intonation".
  • Add "natural emphasis on important details".
  • Add "relaxed conversational pacing".

Pacing and delivery

Pacing is the speed and rhythm of the voice. It affects clarity, personality, and how easy the agent is to understand.

Useful pacing phrases:

  • Speaking normally.
  • Natural conversational pace.
  • Relaxed conversational pacing.
  • Even pacing.
  • Deliberate and measured pacing.
  • Speaking slowly.
  • Speaking quickly.
  • Fast-paced.
  • Calm pace.
  • Staccato delivery.
  • Drawn out rhythm.

For restaurant calls, clear and natural pacing is usually best. You can also tell Otto when the voice should slow down.

Useful examples:

  • Speaks at a relaxed conversational pace.
  • Slows down slightly for phone numbers, addresses, order totals, booking times, and allergy information.
  • Uses gentle emphasis on helpful information.
  • Keeps timing even between words.

Avoid very fast, erratic, or staccato delivery unless that is deliberately part of the character you want.

How to refine your prompt

If the first previews are not right, change one or two parts of the prompt and generate previews again.

Try these adjustments:

  • If the voice feels generic, add age, persona, emotion, accent, and pacing.
  • If the voice sounds too theatrical, remove character-style language and ask for a natural customer-service voice.
  • If the audio sounds rough, remove special effect words and add "excellent quality" or "studio quality".
  • If the accent drifts, move the language and dialect to the first sentence.
  • If the tone is wrong, separate the physical sound from the emotion. For example, "warm, smooth timbre" and "calm, reassuring emotion".
  • If the pace is wrong, use direct wording such as "relaxed conversational pacing" or "deliberate and measured pacing".

Small wording changes can produce different previews. For example, "perfect audio quality" and "the audio quality is perfect" can behave differently. If a prompt is close but not quite right, try rewording rather than rewriting everything.

Example prompts for restaurants

General Australian restaurant host

Use this when you want a clearly Australian-sounding restaurant voice.

Thick Australian accent. Female native Australian English speaker. 20-30. Excellent quality.

Bubbly, energetic local restaurant host. Friendly, confident, helpful. Strong Aussie pronunciation, relaxed Australian rhythm, non-rhotic R sounds, and clear conversational delivery. Avoid all American pronunciation.

Premium dining venue

Native Australian English only. Male, 35-50. Broadcast quality. Persona: polished reservations host. Emotion: calm, professional, reassuring. Deep, smooth, resonant voice with neutral Australian pronunciation and measured pacing. Sounds attentive and composed without sounding stiff or overly formal.

Casual takeaway venue

Native Australian English only. Gender-neutral, young adult. Excellent quality. Persona: upbeat takeaway assistant. Emotion: friendly, energetic, helpful. Light, clear, mid-pitched voice with a natural Australian accent and relaxed conversational delivery. Speaks clearly and keeps the pace efficient without rushing.

Family restaurant

Native Australian English only. Female, 40-55. Excellent quality. Persona: warm family host. Emotion: welcoming, patient, kind. Warm, mellow voice with smooth timbre, neutral Australian pronunciation, and gentle intonation. Speaks at an even pace and slows down slightly for booking times, addresses, and special requests.

 

Restaurant with a subtle accent influence

Native Australian English with a slight Italian accent influence. Male, 35-50. Excellent quality. Persona: welcoming restaurant owner. Emotion: warm, confident, generous. Rich, resonant voice with clear pronunciation, relaxed pacing, and friendly emphasis. Sounds natural for Australian customer calls, with only a subtle accent influence.

Common mistakes to avoid

  • Do not use "accent" when you mean intonation, emphasis, or delivery.
  • Do not use vague accent words such as "foreign" or "exotic".
  • Do not use special effect words such as "reverb", "echo", "phone", or "tape" for a normal restaurant agent.
  • Do not rely on slang to create an Australian accent.
  • Do not mix too many competing traits in one prompt.
  • Do not forget to specify language and dialect in the first sentence.
  • Do not describe the voice only as "nice", "good", or "friendly" if you need a specific result.

What happens next

After you save a custom voice, assign it to the Otto agent that should use it. Place a test call after changing the voice so you can hear how it sounds to customers.

If the voice is close but not right, return to your prompt, adjust the most important detail, and generate new previews.