Product notes and boundaries
A plain account of what the studio does now, what it does not do yet, and how we talk about waiting. · Updated 2026-11-12
VoxMuse is an AI audio studio for independent creators. You sign in with Google. From there you can queue a song from lyrics and a style, or a short spoken Chinese line with a delivery prompt. Voice cloning is only a form and a consent checkbox.
What this release covers
In the product today:
- The landing page and the studio layout
- Google sign-in for studio, works, and related pages
- Music and short voice job forms, validation, and remaining-count feedback from your account
- A clone form that checks consent and does not upload a recording
- My works, with play and download when a job has audio
- Privacy, terms, and these notes
Still limited:
- Speech on the live service is not a finished, guaranteed voice worker. A queued voice job may wait without audio.
- Sound effects are not in the studio yet.
- Voice cloning does not create a voice.
- There is no live or streaming generation.
Jobs are written to your account and shown with a real queue state. The interface never fills the gap with fake percentages or placeholder audio.
Capabilities and boundaries
| Capability | How this release expresses it | Boundary |
|---|---|---|
| Voice | A short line plus a delivery style, queued like music | Not a live stream. Play and download only if speech audio is actually ready. Do not treat the live site as end-to-end voice. |
| Music | Lyrics plus a style, queued as a studio job | Short songs first. Wait time depends on the song worker. |
| Voice cloning | Name, file picker, and consent checkbox | No recording is accepted or stored in this build. |
Task states
Jobs move through queued, running, uploading, and succeeded, with failed, cancelled, and expired as other endings. We show a percentage only when real progress exists. If nothing is producing audio, the studio says it is waiting — it does not invent a result.
Accounts and allotments
Studio use requires a Google account. Free music, voice, and clone counts are one-time per account and do not reset every day. The remaining numbers you see in the studio come from the account service, not from this page.
Next steps
- Measure real speech and short-song latency, success, and cost before promising turnaround.
- Keep music and voice jobs recovering after a refresh, without double-creating work.
- Open voice cloning only with consent, quality checks, and deletion.
- Keep replacing example copy with real work, and keep play and download honest about when audio exists.