Sentioscope

Speech and voice controls, as each vendor documents them

Digital replica: a voice treated as a likeness

A digital replica is a synthesised likeness of a real person, and a voice can be one. One vendor here says so, in a compliance checklist, and names United States statutes alongside the claim. As of 2026-09-12.

One page that restricts, and several that enableA compliance checklist treats an identifiable voice as a likeness question and names United States statutes. API references describe how to enrol a clone and say nothing about permission.Written to restrictWritten to enableThe documentA compliance checklistAn API reference or quick startWhat it saysAn identifiable voice is a likenessquestionHow much audio a clone needsWhat it namesDigital replica statutesA threshold, a format, an idPermissionRaised, and left undrawnNot addressed at allOne route here needs no recording and therefore no person
Fig. 1 A reference recording and a clone of an identifiable voice are not always easy to separate, and the checklist names the risk without drawing the line.
Digital replica, as published. Recorded 2026-09-12.
Where it appearsWhat it saysWhat it leaves open
A compliance checklistAn identifiable voice is a likeness questionWhere the line between reference and clone falls
An API quick startHow to enrol a clone, and from how much audioWhether permission was obtained
A voices referenceHow to build a voice from a sampleWhose sample it may be

Inclusion rule. The kinds of document in this register that touch permission to use somebody's voice, and what each kind says about it. Order. Fixed order: the document that restricts, then the documents that enable.

1A restriction is documentation, and there is only one here

Most of this register records what a vendor says a production may do. One entry records what it says a production may not, and in doing so names the two permitted routes: a preset library and the user's own recordings.

The framing is unusual enough to note. A compliance checklist is written to constrain rather than to sell, so the information in it is probably conservative and definitely incomplete.

2The line itself is left undrawn

A reference recording of a performer and a clone of an identifiable voice are not always easy to separate, particularly where a short clip establishes a voice for one call. The checklist names the risk and does not say where the boundary falls.

This register does not draw it either. Legal questions about likeness are outside what any of these pages settles, and a page that guessed at the boundary would be worse than one that records the silence.

3A described voice sidesteps the problem

One entry lets a voice be asked for in words rather than built from a recording, with a published minimum of twenty characters. There is no identifiable person behind a sentence about a voice, which makes it the cleanest route for a character nobody was going to record.

It is also the least controllable, producing something plausible rather than something chosen. That is a reasonable trade for a character with three lines and a poor one for a lead.

4Sources read for this entry

This page defines the term and logs what each vendor documents about it, each figure read on 2026-09-12. Related: Clip length, On-screen speaker.