ElevenLabs and Mistral have launched AI tools that enable real-time voice translation across multiple languages, with ElevenLabs supporting over 90 languages and Mistral covering 35+ languages. These tools allow a single voice to speak in languages its owner never recorded, raising significant implications for voice actors and businesses.
TL;DR
- ElevenLabs and Mistral have launched AI tools that enable real-time voice translation across multiple languages.
- ElevenLabs' Eleven Realtime Translation supports over 90 languages and is currently in alpha for enterprise customers.
- Mistral's Voxtral 3 can adapt a speaker's voice to 35+ languages using just 30 minutes of audio.
What happened
ElevenLabs announced Eleven Realtime Translation on October 21, 2026. This tool allows for real-time translation during calls, enabling participants to speak in their own language and hear others in theirs. The company targets phone networks, contact centers, and live events, aiming to reduce the need for interpreters. ElevenLabs' tool uses its Eleven v4 Turbo speech model to maintain voice consistency across languages. Additionally, Deutsche Telekom is integrating the translation model into its Magenta AI Call Assistant.
Mistral, a Paris-based AI company, released Voxtral 3 the same week. This family of speech models includes features for transcription, voice agents, and text-to-speech. The Pro Voice feature allows companies to adapt a speaker's voice to 35+ languages using just 30 minutes of audio. Mistral's models can be run through their API or deployed on a company's own infrastructure, with voice agent costs ranging from $0.04 to $0.08 per minute.
Why it matters
These tools have significant implications for voice actors, as they enable the use of a single voice across multiple languages. ElevenLabs' Voice Library, where voice actors can share and license their voice clones, has over 10,400 creators earning from usage. However, the lack of clear consent and acceptable-use frameworks from Mistral raises questions about the ethical use of these tools.
For businesses, these tools offer the potential to provide customer support in multiple languages without the need for multilingual staff. However, the use of these tools may be subject to existing union contracts and licensing agreements, which may require specific consent for language translation and dubbing.
Key facts
- ElevenLabs' Eleven Realtime Translation supports over 90 languages.
- Mistral's Voxtral 3 can adapt a speaker's voice to 35+ languages using just 30 minutes of audio.
- ElevenLabs' Voice Library has over 10,400 creators earning from usage.
- Mistral's voice agent costs range from $0.04 to $0.08 per minute.
- Deutsche Telekom is integrating ElevenLabs' translation model into its Magenta AI Call Assistant.
- SAG-AFTRA's new television animation deal requires specific consent for using digital replicas in language dubbing.
- Japan's Actors Union guidelines require separate permission and fees for AI translation or dubbing of performances.
Context
The launch of these tools comes amid growing concerns about the ethical use of AI in the entertainment and customer service industries. While these tools offer significant benefits in terms of language accessibility and efficiency, they also raise questions about the rights of voice actors and the potential for misuse.
The AI industry is increasingly focused on developing tools that can bridge language barriers. These launches from ElevenLabs and Mistral are part of a broader trend towards real-time translation and voice adaptation technologies. However, the industry must also address the ethical implications of these technologies to ensure they are used responsibly.
