AreaAudit

Which languages each generator says it supports, and where

The language register as a file

The register is published as one file. Each row is a single language claim, tied to the column it belongs in and to the page it came from. As of 2026-09-12.

Why the column an entry belongs to travels with itAn interface claim, an output claim and a voice claim look identical once they are out of context. The file carries the column with every row, so a rebuilt comparison cannot count a translated menu as evidence that a model speaks the language.Easiest claim at the topInterfaceMenus and help. Cheapest to publish, most often counted inheadlines.OutputThe language of the generated script or dialogue.VoiceA named voice able to speak it. Rarest claim, and the one aproduction needs.The file keeps all three apart
Fig. 1 Merging interface and output claims is the commonest error in coverage tables, and it happens in the copy step rather than in the research.

Language lists are the fastest-moving claims on this site and the least often dated. A file with a reading date on every row makes it possible to see a list grow rather than only to see its current state.

The column an entry belongs to is carried in the file, so an interface claim can never be counted as an output claim by whoever rebuilds the comparison.

Rows that record an absence are exported alongside the rest. A market with no claims is the finding a buyer in that market most needs, and dropping those rows would make the file look tidier and mean less.

Reuse is governed by the licence named below. Whatever it requires, the source and date columns are the part worth keeping: a coverage claim without them cannot be re-checked.

1What each column holds

Every column in the register export, left to right. Recorded 2026-09-12.
ColumnWhat it holds
fact_idThe row's permanent handle. Corrections change the value, never this.
valueWhat the vendor's page declares, transcribed rather than summarised.
applies_toWhich of the three language columns the row belongs to.
source_urlThe exact page the declaration was read on.
source_nameHow the vendor titles that page.
value_sinceThe earliest date this site can show the statement held.
verifiedWhen the page was last opened to check it still says this.

Inclusion rule. Every recorded claim is exported, including those recording that a vendor names nothing. Order. File order follows recording order, grouping rows by tool.

The file is at /datasets/languages.csv, licensed ODbL 1.0. Cite the register and the reading date.

The rest of the atlas: Columns, Entries, By language, Side by side, Markets, Tools, Output, Questions, Terms, Learn. How a list earns a column is set out on the counting page.

2Languages in a working pipeline

On the language side of “Data”, SceneMixer's published guides say the following:

The spoken lines and the end card follow your interface language across fifteen dialogue languages, with the cast and setting adapted to that market. The end card falls back to a language it can typeset when a script needs shaping or right-to-left layout.

SceneMixer, AI Ad Video Generator — “Which languages can the ad be in?” (read 2026-10-13)

SceneMixer separates the script language from the production language: the output (episode outlines, storyboards, on-screen text) follows your interface language, and target-region localization adapts character names, casting, and scenery for Western, East Asian, Southeast Asian, or Latin American audiences.

SceneMixer, Turn a Novel into a Video Series with AI — “Can I turn a Chinese web novel into an English-language drama?” (read 2026-10-13)

The Quote column. Each reading in the file also carries the words on its source page that it was read from, so any row can be checked against that page. Rows recorded as not stated on the pages read leave it empty.