ElevenLabs drops AI Dubbing that speaks your voice in 20+ languages

Watching translated videos usually comes with a frustrating trade-off: you either read wall-to-wall subtitles, or listen to a generic voice actor who sounds nothing like the original speaker.

Voice AI specialist ElevenLabs is setting out to fix that disconnect. On October 10, the company officially launched AI Dubbing, a tool that can translate audio and video into different languages within minutes while preserving the original speaker’s distinct voice, tone, and delivery.

Multilingual cloning with human nuance

Most automatic translation tools strip away the personality of the speaker. ElevenLabs is taking a different route by leveraging its Eleven Multilingual v2 model, supporting voice translation across more than 20 languages, including Hindi, Portuguese, Spanish, Japanese, Ukrainian, Polish, and Arabic.

What makes this launch stand out is how it handles the subtleties of speech. The system is designed to retain the original speaker's emotional range and intonation rather than defaulting to flat text-to-speech output. If a creator sounds excited or hesitant in English, that performance is meant to carry over directly into Spanish or Japanese.

Handling messy real-world media is another major hurdle the company tackled. ElevenLabs built dedicated mechanisms to separate dialogue from background music and environmental noise. It also detects multiple distinct speakers in a single clip, ensuring that a multi-person podcast or interview does not end up with voices scrambled together.

This release builds directly on top of ElevenLabs Studio, the company's recently launched workflow for long-form audio editing, which already introduced speaker management, pause controls, and segment regeneration.

Why it matters

For independent creators and educators, this dramatically rewrites the economics of global reach. As ElevenLabs points out, native English speakers make up just 6% of the global population, and 75% of people do not speak English at all. Translating video into multiple foreign languages used to require expensive agency contracts and weeks of coordination; now it can be tested in minutes for a fraction of the cost.

For the translation and dubbing industry, this adds intense pressure on low-to-mid-tier voiceover work. When an AI can accurately replicate a creator's own voice across twenty languages in minutes, paying human studios for standard localization will quickly become hard to justify.

We think ElevenLabs made a smart move by integrating noise removal, speaker diarization, and emotion preservation into a single self-serve flow. However, the obvious missing piece here is video lip-sync. A translated voice that sounds identical to the speaker is impressive, but if their mouth movements are wildly out of sync on screen, viewers will still experience that uncanny valley. Bringing this audio tech together with synchronized visual synthesis is likely where the real platform war will happen next.