AI Trends

OpenAI Explains in Detail How to Configure GPT-Live for Multiple Languages and Regional Accents

OpenAI DevelopersUpdated Nov 20 04:11

AI summary

OpenAI explains how built-in voices, prompt-to-voice, script_hint, and session instructions can make GPT-Live better suited to different languages and regional accents. Using Indian English, Castilian Spanish, and Dubai Arabic as examples, the article recommends testing voices with real conversations, numbers, proper nouns, language switching, and correction scenarios. For generated voices, developers must also obtain the speaker’s explicit consent, disclose that the voice is AI-generated, and note that voice creation is not available with Zero Data Retention.

Why read this

SelectedAI score 69

The article breaks regional accent design down into voice selection, session prompts, and real-world scenario testing, an approach that can be applied to multilingual voice agent development.

Topics

Tags

Body · AI translation

Even if a voice agent speaks Spanish, it can still sound off to the audience you’re building it for. It might use an accent from a different region, choose wording that feels too formal, or read a price in a way a local speaker wouldn’t expect.

While GPT-Live’s default options can go a long way, there are additional considerations when building voice applications for audiences across languages and regions. Let’s look at voice selection and session instructions for achieving better results, using regional varieties such as Indian English, Spanish from Spain, and Emirati Arabic.

Start with the right voice

First, check GPT-Live’s built-in voices. Some may already have a linguistic or regional influence close to what you need: for example, Bossa and Tempo have a Brazilian Portuguese influence, while Nira is a Hindi voice. Test them with your own conversations before designing something new, as the built-in voices offer high naturalness out of the box.

Find a built-in voice Additional GPT-Live voices by language, regional influence, and presentation

Language / regional influenceFeminineMasculine

Australian English Quartz (generated source) Ripple (natural source)

British English — Vesper (natural source)

Irish English Willow (natural source) Stone (natural source)

North American English Gleam (natural source) Meridian (natural source)

Filipino English — Beacon (generated source)

Southern U.S. English Delta (generated source) Cinder (generated source)

Brazilian Portuguese Bossa (natural source) Tempo (natural source)

German Nuri (natural source) Juni (natural source) Flitz (natural source)

French Brise (natural source) Sillage (natural source)

Hindi Nira (natural source) —

Japanese Harema (natural source) Shida (natural source)

Korean Noeul (natural source) — Natural source Generated source

Use lowercase API names when selecting a built-in voice (for example, nira). The voice options guide contains the complete current catalog.

If you need a more specific regional sound, you can use prompt-to-voice to generate a synthetic voice from a text description. Describe the region as precisely as possible, along with the pitch, pace, and vocal character you want. When generating a voice, you can use the script_hint field to add representative speech in the target language. For a bilingual voice, include both languages in the sample text and describe the desired sound in each. Use original adult voices, and do not impersonate real people. See the responsible-use guidance.

Eligible customers can also create a custom voice from a recording of a consenting speaker. Check the supported models and accents before choosing a recording-based voice. Before uploading, obtain the speaker’s express consent and secure the rights needed to use the recordings and resulting voice in your application. The consent recording and reference sample must come from the same speaker. For any generated voice, make it clear to listeners that they are hearing AI-generated speech.

Voice creation is not eligible for Zero Data Retention. See the data controls guide for the settings that apply to GPT-Live sessions.

Example: Hindi-influenced English with Nira

For an English-speaking agent serving users in India, set session.audio.output.voice to "nira" when creating the session, then add this prompt to the session instructions:

You are having a natural, clear English voice conversation. Speak fluent English in your natural Hindi-influenced Indian English accent. Maintain that familiar rhythm throughout.

Example: Spanish from central and northern Spain

Suppose you’re building a shopping assistant for customers in Spain. Instead of asking for a generic Spanish accent, open the Voice Playground and try this description in its prompt-to-voice workflow:

An original adult feminine voice, clear and lightly bright, with a medium pitch and gentle warmth. Speak Spanish with a natural standard Peninsular Castilian accent from central or northern Spain, using native Spanish vowels, consonants, stress, and sentence melody. Naturally distinguish s from z and c before e or i, without exaggeration. Use a moderately lively conversational pace, natural pauses, and varied intonation.

Pair it with a Spanish script_hint that gives the voice some ordinary connected speech to work with:

Hello, how are you? I’m here to help you. We can calmly review the options and choose the one that suits you best. Would you prefer to start with the colors or the sizes? This blue jacket goes well with a comfortable pair of shoes. Move closer to the window to see the color in natural light.

Replica of the Voice Playground editor using the Spanish example.

This targets a particular variety of Spanish; the distinction described in the prompt is not universal across Spain or the Spanish-speaking world. With this prompt, you can generate a few candidates, keep the saved voice you prefer, and test it on a new sentence that was not in the creation script until you are satisfied.

Example: Emirati Arabic from Dubai

Arabic is another case where naming the language alone leaves a great deal unspecified. For a conversational Dubai voice, an Arabic voice-design description might be:

An original, fictional adult male voice with a contemporary urban Emirati accent from Dubai. The pronunciation, speech patterns, and rhythm are locally authentic and consistent, while the words remain clear within connected everyday speech. A medium pitch leaning low, with a warm, clear resonance. He speaks at an unhurried pace, with each sentence flowing as a single thought and varied intonation that follows the meaning.

The description asks for an original adult masculine voice with a contemporary urban Dubai accent, warm resonance, a slightly lower pitch, and connected, relaxed speech. Use a conversational Arabic script_hint as well:

Yeah, I saw your message. I’m here by the park—no, I mean by the parking lot at the second entrance. Okay? It’s quieter here than where we were yesterday. Take your time; I’ve only just arrived, and I’m not in a hurry. Do you want us to walk around a bit or sit down? Either is fine, whatever makes you comfortable.

In general, keep the voice design focused on sound. Managing details like whether the agent should answer an Egyptian caller in Emirati Arabic, switch into English, or confirm an order belongs in the session instructions, which we’ll cover later on.

Test the voice on what your application will actually say

But having a solid voice is only the starting point. Once you have a voice, you can build a small listening set with fresh sentences, realistic conversations, and the words or values your application handles most often to stress-test your voice.

A few edge cases are particularly useful to test:

English text is being prepared.

Regional number conventions: The Indian system of numeration uses lakh and crore. One lakh is 1,00,000 (100,000), and one crore is 1,00,00,000 (10,000,000). For an English-speaking agent serving India, test whether ₹1,25,000 is spoken as “one lakh twenty-five thousand rupees” when that’s the intended convention.

Identifiers and corrections: Have a caller spell a fictional booking code, pause, and correct one character. Check that the agent waits for the correction and reads back the final code accurately. Test email addresses and other alphanumeric inputs your product accepts.

Mixed-language speech: Try an Arabic sentence containing an English product name and a price, then switch languages for the next turn. Check the spoken amount, currency, and product name separately from the accent.

For dates, amounts, and identifiers, validate what your application actually sends to the backend as well as what the caller hears. In many regions, 04/05 could mean two different dates, so it may be worth asking for clarification and reading the month back by name.

Prompt the agent for the conversation

Once you’ve chosen a voice, select it when creating the GPT-Live session. For a saved generated voice, set session.audio.output.voice.id to the returned voice ID. Put the agent’s role, language policy, delivery, and conversation behavior in instructions, and reinforce the original voice description there.

For a single-language application, write the instructions and examples in the language you want the agent to speak. Here’s a starting point for the Spanish shopping assistant:

You are a shopping assistant. Speak in Spanish from Spain, with the Castilian accent of the central and northern peninsula, a conversational pace, and natural intonation. Use a friendly and respectful tone without sounding like an advertisement.

Respond in Spanish unless the customer asks to change languages. Ask one question at a time. If a name or size is unclear, ask the customer to repeat it. Do not confirm prices, availability, or changes to an order until the system has verified them.

For a multilingual application, decide what should trigger a language switch and whether the underlying regional speaking style should remain fixed. Those are separate choices. For an Emirati Arabic and English assistant, first extend the earlier voice description with English delivery guidance, such as “In English, speak naturally with a light Emirati influence while keeping the same warmth and lower pitch.” Include representative English speech alongside the Arabic in script_hint, such as “I can check the available sizes for you,” before generating the bilingual voice. Then use a session policy like this:

Answer in the language of the caller's latest clear utterance. Switch between Arabic and English when the caller clearly switches. A short “OK,” number, name, or English product name inside an Arabic sentence does not by itself change the conversation language.

When speaking Arabic, use conversational urban Emirati Arabic from Dubai, even when the caller speaks another Arabic variety. Prefer natural local wording where appropriate: شو for “what,” الحين for “now,” وين for “where,” and مب for negation. Keep the dialect coherent without exaggeration. In English, keep the same speaker identity.

Use these as phrasing examples, not fixed answers: - Clarification: “Do you mean the small box or the large one? And how many do you want?” - Unclear name: “The name isn’t clear to me. Could you repeat it or spell it for me?”

Say amounts in the current conversation language and preserve their exact value and currency. Only confirm a lookup, price, or completed action when the backend supports that statement.

This example deliberately keeps the assistant’s Arabic dialect fixed while allowing switches to English. If your product should accommodate a caller’s requested dialect instead, write and test that policy explicitly. While voice descriptions and session prompts can steer an accent, they do not guarantee that every reply will retain it, so test switches in both directions, including after a session reconnects.

Give examples of the register you want

Language and accent also leave the level of formality open. In Hindi, a courteous support agent might use आप and respectful verb forms. You can make that preference concrete in the session prompt:

Address the customer as “आप.” Use respectful language and verb forms. Say “जी, बताइए” instead of “हाँ, बताओ.” Say “आप कैसे हैं?” instead of “तुम कैसे हो?”

The examples distinguish a casual “yes, tell me” from a more courteous equivalent. That gives the model a clearer target than simply asking it to sound professional. Use phrases a local support representative would actually say, and test them in context rather than translating an English script word for word.

For more on pronunciation cues, personality, greetings, and interruption behavior, see the GPT-Live prompting guide. The voice agent evaluation guide covers how to turn your listening cases into a repeatable evaluation as you change voices, prompts, or models.

Source

OpenAI Developers