Skip to content

Eleven Realtime Translation: our live voice translation model

ElevenLabs

The live voice translation model supports over 90 languages with automatic detection and is available in alpha for ElevenAPI enterprise customers.

Key points

  • Runs on the Eleven v4 Turbo low-latency speech model
  • Supports custom voice clones as well as over 17,500 library voices that stay consistent across languages
  • Lets developers tune between translation latency and output quality depending on the use case
  • Deutsche Telekom is integrating the model into its Magenta AI Call Assistant

Why it matters

Contact centers and telecom operators can conduct live, bilingual phone conversations using consistent synthetic voices without hiring human interpreters.

Sources