xAI appears to be taking Grok’s voice capabilities to a new level with an upcoming Voice Cloning feature spotted in development for its iOS app. The feature would let users record their own voice by reading through a specific text prompt, structured as four paragraphs of a short story, after which Grok captures and processes the recording to generate a personal voice profile. That cloned voice could then be used within Grok’s existing voice mode or shared with others via a link. When someone opens a shared voice link, they would also have the option to save it for their own use.
This moves xAI well beyond simply offering preset voice personalities. Since we first tracked Grok’s voice mode development back in early 2025, when the company was building out options like “unhinged” and “genius” with ElevenLabs’ technology, it has been clear that xAI views audio as a core differentiator. The company has since launched a Grok Voice Agent API, positioned as the fastest on the market with sub-700ms response times, and it now powers experiences in Tesla vehicles as well. Voice Cloning would be a natural extension of that investment, shifting the product from curated presets toward truly personalized audio.
The sharing mechanic is particularly worth watching. A link-based distribution model could organically seed an informal voices marketplace (speculation), something this publication anticipated as platforms began rolling out custom voice options in previous years.
Initially, the most likely beneficiaries would be creators, podcasters, and developers seeking a consistent audio identity across AI-powered products. The feature is still in development, with no confirmed release window, though xAI’s recent shipping cadence for Grok suggests it may not be far off.
