Anthropic appears to be preparing one of the larger updates to Claude’s mobile voice mode since its beta debut last May, with several changes surfacing in the app ahead of any official word. The refreshed UI introduces a new glow animation around the voice orb and a push-to-talk option, a meaningful tweak given that Claude’s current voice flow already follows a turn-based pattern rather than the full-duplex streaming used by ChatGPT’s Advanced Voice or Gemini Live.
The more consequential addition is a new “Language” setting, tagged as beta. English remains the only option live today, but the menu already lists German, Portuguese, Chinese, Japanese, Russian, Ukrainian, and other languages. Most new languages come with two voices each, while some have just one, in contrast to English, which ships with five personas (Mellow, Airy, Buttery, Glassy, and Rounded).
What makes this more than a routine localization push is how the switching works. Beyond the manual selector, users can simply ask Claude to change the language mid-conversation, and it will switch on the fly, a behavior not possible in the existing build. The output still carries the cadence of text-to-speech rather than that of a native speech-to-speech model, hinting at a new orchestration layer managing multiple voices and language profiles behind the scenes, rather than a wholesale move to an in-house audio stack.
That detail matters because Anthropic has so far relied on external providers for Claude’s spoken side, with ElevenLabs listed as a text-to-speech subcontractor and a broader relationship with Amazon powering Alexa+. A multilingual rollout layered on top of that stack would let the company close one of the more visible gaps with rivals, both ChatGPT and Gemini have offered multilingual voice for some time. No timeline has surfaced, and the language list could still shift before the public flip.