Sentioscope

Speech and voice controls, as each vendor documents them

Language code: a language written for software

A language code is a short machine-readable token for a language. One entry publishes one per voice, alongside a formal and a native name. Everywhere else the language answer is a name, a number or nothing. As of 2026-09-12.

Four grades of language answer, only one checkableA code can be matched against a distribution plan automatically. A list of names has to be read and interpreted. A total says the capability is broad. An unnamed claim says it is intended.What software can verifyA code per voiceMatched automatically, and auditable as a catalogue changes.A list of namesCheckable by hand, and ambiguous between market, script anddialect.A totalComparable only where two vendors are counting the samequantity.An unnamed claimCannot be checked, and cannot be wrong, because nothing isnamed.What only a listener can
Fig. 1 Grading a checkable answer above an impressive one is why this register keeps wording rather than scores.
Language code, as published. Recorded 2026-09-12.
How the answer is publishedWhether software can check itWhat it settles
A code per voiceYes, by matching against a planWhich voices reach which markets
A list of namesOnly by hand, and ambiguouslyWhether the next market is in scope
A total with no namesNoThat the capability is broad
An unnamed claimNoThat the capability is intended

Inclusion rule. The grades of language answer found across this register, ordered by how far each can be verified without asking the vendor. Order. Fixed order, from the most checkable to the least.

1Ambiguity is what ruins most language claims

A name in prose can refer to a market, a writing system or a regional variety, and which one is meant is usually decided by whoever reads it. A code removes that decision, which is why it sits at the top of this scale even where no total is published.

It also makes an inventory auditable. A code that disappears is visible to a script; a voice quietly replaced behind a sentence about broad support is visible to nobody.

2A code is worth more than a bigger number

The largest language figure in this register is ninety-five or more, published as a total by a product with no voice inventory. A catalogue publishing a code on each of a smaller set of voices answers a commissioning question that the larger number cannot touch.

Grading a checkable answer above an impressive one is the whole reason this register keeps wording instead of scores. A score would have to choose a number, and the numbers here are not measuring the same thing.

3Interface languages and dialogue languages are different lists

Product pages routinely carry a language list belonging to the menus, the documentation or a prompt parser, and that list is longer than the one the actors speak. Reading the two as one is the commonest way this field gets misread.

A code helps only if it is attached to a voice. A code attached to an interface locale is precise about the wrong thing, and this register records only languages named for the spoken performance.

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: Guide track, Dubbing pass.