ElevenLabs launches Dubbing v2, which it says preserves the original speaker's emotion, tone, and pacing across 90+ languages while staying synced to content
ElevenLabs
Context & Ripple Effects
ElevenLabs’ recent coverage shows it expanding from voice generation into a broader audio stack: Music v2 was positioned for commercially cleared music creation, while ElevenReader added a large licensed audiobook catalog as the company sought a larger consumer audio role.
Dubbing v2 extends that stack into localization. Its claimed ability to retain performance characteristics while matching the source timing targets a key limitation of automated translation for video and spoken-content publishers.
First-order effects
Video creators, publishers, and media platforms can use ElevenLabs’ new dubbing product to produce localized versions across more languages while keeping dialogue aligned with the original content.
ElevenLabs gains a more complete production offering alongside its music, sound-effects, and audiobook products, rather than serving only as a voice-generation vendor.
Second-order effects
Localization providers and competing AI voice tools face pressure to match not just translation coverage but performance preservation and synchronization, the qualities most relevant to finished video.
The product can make multilingual distribution more practical for owners of existing spoken-content libraries, increasing the value of workflows that combine rights-cleared source material, translation, voice, and final audio production.
Third-order effects
If quality claims hold in production use, AI audio competition may shift from single-purpose generation models toward integrated, rights-aware content pipelines spanning creation, localization, and distribution.
The expansion also raises the stakes for consent, licensing, and control over voice performance, particularly as tools aim to reproduce expressive attributes rather than merely translate words.
The trend: Dubbing v2 is part of the shift from standalone generative-audio tools toward end-to-end AI systems for producing and distributing localized media.
ElevenLabs just launched Dubbing v2 Alpha. The model can translate speech across languages while preserving the speaker's emotion, tone, and delivery. Content that once felt lost in translation can now feel native anywhere. A major step forward for global creators.
ElevenLabs just dropped Dubbing v2. 🤯 This isn't another flat AI voice swap. For the first time, the original actor's emotion, timing, tone, and performance actually survive the translation. The model conditions directly on the real delivery — not a sterile transcript.
ElevenLabs just shipped Dubbing v2. For the first time, the emotion of the original audio carries into every language. Not just the words. The tone. The pauses. The pain. The laugh. 90 plus languages and accents supported. One click. The geographic ceiling on creator income
ElevenLabs launched Dubbing v2, a new AI dubbing model designed to preserve the emotion, tone, and performance of the original content across 90+ languages. Dubbing v2 is built to solve one of the biggest problems in AI dubbing: flat, unnatural audio that loses the original [vide…
A little over a year ago @MrBeast told Zuck on @ColinandSamir that broken dubbing, which doesn't carry well the culture, vibe and meaning from English to Spanish is the biggest barrier to his growth as a creator on Meta. Creator distribution just got an incredible upgrade.
Introducing @ElevenLabs Dubbing V2, the world's SOTA model for high-quality dubbing. ▪️ This is an Audio-to-Audio model that preserves emotion & intent ▪️ 90+ languages supported ▪️ Available through ElevenCreative with API coming soon Kudos to our entire Research & Engineering […
🤯 Language barriers? Where we're going we won't have any language barriers Dubbing v2 by @ElevenLabs blew my mind so much I had to stop editing @thursdai_pod and show you (100% not paid lol) it's insane, it understands tonality, emotion, does voice transfer & even accent!? [video…
Introducing Dubbing v2. For the first time, AI dubbing preserves how something was said, not just what was said. Dubbing v2 reads the original audio directly rather than just the transcript, so your emotion, tone, and delivery carries across 100+ languages. Every system before [v…
Introducing Dubbing v2, our revolutionary new dubbing model. For the first time, the emotion and performance of the original content is carried over into every language. [video]
Dubbing v2 is now live! This is a new type of architecture - a fully end-to-end dubbing model. By conditioning on the original audio, it's able to carry over the original emotions and performance [video]
Dubbing is officially dead. 🔥 ElevenLabs just shipped Dubbing v2 and it solves the one problem AI dubbing could never crack. Every dubbed video until today sounded flat. Robotic. Dead. Because every tool translated the transcript, not the performance. Dubbing v2 conditions