How we curated ElevenLabs voices for news publishing

This post is narrated by Caroline Piercy's Instant voice clone, powered by ElevenLabs through BeyondWords.

ElevenLabs offers more than 10,000 AI voices across a range of styles, tones, and personalities. 

If you’re looking for pre-made voices to narrate your articles, the abundance of choice quickly becomes friction.

After all, not every voice lives up to newsroom standards.

That’s why, when adding ElevenLabs voices into BeyondWords, we didn’t just add the full library—we handpicked the very best options for news narration.

Key criteria for news narration 

First, we excluded ElevenLabs voices that use “Live Moderation”. 

News publishers regularly cover sensitive or controversial topics, and automated moderation systems aren’t always calibrated for journalism. By removing moderated voices from our selection, we help ensure legitimate reporting isn’t inadvertently flagged or restricted.

Secondly, we filtered out voices with short notice periods to give publishers greater stability once a voice is selected.

The final filter we applied was for “High-Quality” voices. 

These voices have been reviewed by the ElevenLabs team to meet professional standards for clarity, stability, and tone—all essential to doing publishers’ journalism justice.

Once we’d applied our initial filters, it was time to start listening.

Curation through careful listening

Our team listened to hundreds of remaining ElevenLabs voices to assess their suitability for article narration, testing them with long-form news content.

We evaluated voices against the qualities publishers consistently tell us matter most: naturalness, clarity, and authority.

This meant removing: 

• character voices, like those made for cartoons;
• novelty voices, like those created for ASMR content;
• voices that didn’t meet our studio-level quality standards; and
• voices with overly emotional or flat delivery.

Where we didn’t have in-house language expertise, we brought in native speakers to share their opinions on the voices. We also gathered feedback from some of our publishers.

The result is a selection of over 200 news-ready voices across dozens of languages and accents*. All available for immediate use through the BeyondWords platform.

Many of these voices are multilingual, which means they're capable of delivering natural narration across various languages. This is particularly useful for publishers who want to maintain consistency across markets.

Multilingual voices can carry traces of their native accent, but they’re often indistinguishable from native voices. It’s worth exploring the full range before narrowing your selection.

Tools for faster voice selection

To make voice selection even easier, we added voice previews built around real news-style openings. These give you a realistic sense of how each voice will perform in context.

We also set new default voices to give you a convenient starting point across our most-used languages.

You can preview some of our favorite ElevenLabs voices below: 

If you’d like to get tailored voice advice from our team, we’re happy to help.

Finding the right voices for every publisher

We have vast experience working with publishers to find the right voices for their brands.

Just recently, we helped a fashion title adopt a female, regionally accented voice that feels authentic to its readership and aligns seamlessly with its style.

Whether you have a detailed brief or prefer to rely on instinct, we’ll guide you toward voices that align with your editorial identity. And collaborate to find the right fit for every use case.

If there’s a specific ElevenLabs voice you’d like to use that isn’t part of our curated selection, we’ll add it for you. You can also create ElevenLabs voice clones.

Ready to find your publication’s voice? Book a demo with our team.

*BeyondWords currently offers ElevenLabs voices for Afrikaans, Arabic, Armenian, Assamese, Azerbaijani, Belarusian, Bengali, Bosnian, Bulgarian, Catalan, Cebuano, Chichewa, Croatian, Czech, Danish, Dutch, English, Estonian, Filipino, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Hausa, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Irish, Italian, Japanese, Javanese, Kannada, Kazakh, Kirghiz, Korean, Latvian, Lingala, Lithuanian, Luxembourgish, Macedonian, Malay, Malayalam, Mandarin Chinese, Marathi, Nepali, Norwegian, Pashto, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Serbian, Sindhi, Slovak, Slovenian, Somali, Spanish, Swahili, Swedish, Tamil, Telugu, Thai, Turkish, Ukrainian, Urdu, Vietnamese, and Welsh.

Visit our docs for a full list of supported languages and accents.