Status First
Voice cloning on x-rush.ai is a designed, specified capability of Voice Studio, and as of October 2026 it is not live: there is no cloning in production, no production voice-provider integration, and no launch date to print. The Audio module's live feature today is Music Gen. This page documents the frozen design so you can plan around it, and it will say so plainly if you read it after launch; check the Voice Studio page for the current status before building a workflow that depends on it.
How Voice Cloning Is Designed to Work
The frozen flow is deliberately boring:
- Upload samples — before a task starts, the interface shows which audio formats, sample counts, lengths, and quality levels are currently supported. Those requirements are maintained in the official product contract rather than fixed in a marketing page, so this page quotes none.
- Analysis — pitch, timbre, rhythm, and texture are analyzed to build a voice.
- A private voice is created — the cloned voice lands in the current Project's voice library, next to the built-in ones.
- Text-to-speech — narration runs as whole-body text against a single selected voice; the design has no multi-speaker tracks, sentence-level editing, or partial redo inside one generation.
Where a Clone Can Go
Nowhere far, by design. A clone lives inside the Project where it was made: its samples, configuration, and the voice itself. There is no global account-wide voice library and no community publishing for cloned voices. If you move a workspace to another Project, the confirmation screen strips the voice assignment; the script travels, the voice does not. Deleting a clone destroys its samples and configuration permanently, and the deletion is not refunded.
The Voice Library Alternative
If you don't need a specific person's voice, Voice Studio's library ships with a first batch of built-in voices that are licensed for product use and passed quality review. Celebrity imitation is not part of the lineup. Each voice's page shows what it actually supports, language, accent, scene fit, and its real controls (emotion, style, and speed, plus pauses, pronunciation, and pitch where the voice supports them), rather than a universal feature list. See the voice library page for the full picture.
Consent and Ethical Guidelines
Consent is a hard gate, not a terms-page clause:
- You must have the rights to the voice being cloned.
- Cloning without consent violates the terms of service.
- Misuse can end service access.
- Suspected abuse can be reported for review.
Public audio passes a pre-publish safety check and carries the platform watermark on public copies; private results stay private. What the spec deliberately does not define is round-the-clock misuse surveillance, so this page won't claim one.