HomeAI AudioModelsElevenLabs Multilingual v2
One voice, 29 languages

ElevenLabs Multilingual v2 in Pika: Speech Guide

ElevenLabs Multilingual v2 turns scripts into natural speech in 29 languages, and can keep the same voice across all of them. This guide covers when to use it, how to write scripts that sound right, and how to localise a video into several languages.

Pika Soundtrack preview artwork
Audio artwork: Pika.

What is ElevenLabs Multilingual v2?

“Elevenlabs Multilingual v2 is a text-to-speech model with support for 29 languages. Especially useful for content localization, you can upload an audio sample to maintain the speaker’s voice characteristics across all languages.”

Source: Pika’s ElevenLabs Multilingual v2 page.

ElevenLabs describes it as “our most advanced, emotionally-aware speech synthesis model,” producing “natural, lifelike speech with high emotional range and contextual understanding across multiple languages.” It “delivers consistent voice quality and personality across all supported languages while maintaining the speaker’s unique characteristics and accent.”

Source: ElevenLabs models documentation.

The 29 supported languages

ElevenLabs lists these for its Multilingual v2 model:

English (USA, UK, Australia, Canada) Japanese Chinese German Hindi French (France, Canada) Korean Portuguese (Brazil, Portugal) Italian Spanish (Spain, Mexico) Indonesian Dutch Turkish Filipino Polish Swedish Bulgarian Romanian Arabic (Saudi Arabia, UAE) Czech Greek Finnish Croatian Malay Slovak Danish Tamil Ukrainian Russian

Source: ElevenLabs models documentation. Check the options shown in Pika’s app.

Best uses

ElevenLabs says the model “excels in scenarios requiring high-quality, emotionally nuanced speech”:

Character voiceovers

“Ideal for gaming and animation due to its emotional range.”

Professional content

“Well-suited for corporate videos and e-learning materials.”

Multilingual projects

“Maintains consistent voice quality across language switches.”

Stable quality

“Produces consistent, high-quality audio output.”

The trade-off, according to ElevenLabs: it “has a higher latency & cost per character than Flash models,” but “delivers superior quality for projects where lifelike speech is important.”

Multilingual v2 or another Pika speech model?

Pika’s speech models, as described on their Pika pages
Model Pika’s description Choose it when
ElevenLabs Multilingual v2 29 languages; upload a sample to keep a voice across languages Localising content, or when one voice must speak several languages
ElevenLabs Turbo v2.5 “High-speed, high-quality,” 32 languages Speed matters, or you need a language outside the 29
Pika Speech “The fastest, most cost-efficient text-to-speech model on the market”; 76 voice presets; 15,000 characters per request; 48 kHz Long narration with a preset voice, on a budget

Sources: Pika pages for Multilingual v2, Turbo v2.5 and Pika Speech. “Choose it when” is our suggestion.

How to use Multilingual v2 in Pika

  1. Open ElevenLabs Multilingual v2 at create.pika.art/apps/eleven-multilingual and sign in through Pika’s login page.
  2. Paste your script in the language you want spoken.
  3. Choose a voice, or upload an audio sample of a voice you have permission to use.
  4. Check the credit cost the app shows, then generate.
  5. Listen with headphones for pronunciation, pacing and emotion.
  6. Fix the script, not the audio: change punctuation or spelling and regenerate.
  7. Download and add it to your video.

Pika’s page doesn’t list prices or script length limits; the app shows them. See Pika pricing.

Writing scripts that sound natural

  • Write for the ear: short sentences, one idea each.
  • Use punctuation to pace: commas for short pauses, full stops for longer ones.
  • Spell out tricky items: “twenty twenty-six” rather than “2026” if the reading matters; spell out abbreviations the way they should be said.
  • Write names phonetically if they’re mispronounced, then adjust until they sound right.
  • Use the target language’s own script and accents, rather than transliteration.
  • Break long scripts into scenes so you can redo one part without regenerating everything.
Example script: Every morning, before the city wakes, the harbour comes alive. The boats go out one by one. And the gulls follow, as they have for a hundred years.

Localising a video into several languages

1 Script Final version in one language
2 Translate Checked by a fluent speaker
3 Generate Same voice, each language
4 Publish One video per language
  1. Lock the picture first. Finish the video so timings don’t change.
  2. Translate the script and have a fluent speaker check it. Translations often run longer or shorter than the original.
  3. Use the same voice for every language, choosing it or uploading a permitted sample, so the video keeps one identity.
  4. Generate each language and time it against the picture.
  5. Adjust the script, not the video, if a translation runs long.
  6. Mix with music from Pika Music or MiniMax Music 3.0 and export one version per language.

The same line in three languages

English: Welcome. In the next two minutes, we’ll show you how it works.
Spanish: Bienvenidos. En los próximos dos minutos, les mostraremos cómo funciona.
French: Bienvenue. Dans les deux prochaines minutes, nous allons vous montrer comment ça marche.

Example script and translations are our own; have any translation checked before publishing.

Common problems and fixes

Troubleshooting Multilingual v2
Problem Fix
Name or word mispronounced Spell it phonetically in the script
Reading sounds rushed Add commas and full stops; shorten sentences
Numbers read oddly Write them out as words the way they should be said
Wrong accent or language Check the script is fully in the target language, with its accents
Translated version too long for the video Shorten the translated script rather than speeding up the audio
Uploaded voice doesn’t sound right Use a clean, clear sample with no music or background noise

Multilingual v2 FAQs

How many languages does it support?

29, according to both Pika and ElevenLabs. The full list is above.

Can I use my own voice?

Pika’s page says you can upload an audio sample to keep a speaker’s voice across languages. Only upload your own voice or one you have permission to use.

Is it faster than Pika Speech?

Pika calls Pika Speech “the fastest, most cost-efficient text-to-speech model on the market,” and ElevenLabs notes Multilingual v2 has higher latency than its faster models. Choose Multilingual v2 for quality and multiple languages.

Can it sing?

It’s a speech model. For songs with vocals, use MiniMax Music 3.0 or Pika Music.

Is this the official Pika or ElevenLabs website?

No. pikaais.com is an independent informational guide, not affiliated with Pika or ElevenLabs.

Voice one line in three languages

Pick a voice, generate the example line in English, Spanish and French, and listen for how the voice carries across.

Open Multilingual v2 ↗

Next: Voiceovers · Pika Speech · AI audio hub

Sources: Pika pages for ElevenLabs Multilingual v2, ElevenLabs Turbo v2.5 and Pika Speech, ElevenLabs models documentation and Pika pricing, checked September 24, 2026. Image: Pika. Scripts, workflow and tips are our own. Not affiliated with ElevenLabs.