Twenty-seven locales against two, in the same column
CapCut declares 27 same-host site alternates and D-ID declares two. They are the widest and narrowest entries in this column, and neither figure predicts what the other two columns hold for those vendors. As of 2026-09-22.
| Question | CapCut | D-ID |
|---|---|---|
| Site language | 27 same-host alternates | Two: German and English |
| Code form | Region-tagged but for English | Both bare |
| Voice languages | No list published | More than 120 claimed, none named |
| Output languages | No list published | No list published |
Inclusion rule. Two vendors chosen because they hold the extremes of one column. Rows are the register columns plus the code form, which is what the extremes differ in. Order. Fixed row order, widest declared list in the left-hand column.
1The gap is thirteenfold and means less than it looks
Twenty-seven against two is the largest ratio anywhere in this register, and it is a ratio between marketing budgets. Translating a product site is string work, and it scales with how many markets a company already sells into.
Read the other rows and the ordering reverses. One vendor publishes a spoken-language claim and the other publishes none, so on the question a production actually asks, the vendor with two site languages is the one that has said anything at all.
2Two vendors, two different silences
One says nothing about voices while publishing a count of the languages its subtitles can be produced in. The other says nothing about the documents it generates while claiming more than 120 spoken languages without naming one of them.
Both silences are recorded as silences. The difference worth noting is that one sits next to a number and the other next to nothing, and a number with no list behind it cannot be checked against any particular market.
3What would settle the comparison
A named list from either side. If the 120 were published, two site alternates would stop being the interesting part of that row; if the voices behind the subtitle count were published, 27 locales would look like the smaller claim.
Until one of them does, the only honest summary is that these two vendors have spent their disclosure in opposite places, and that a reader comparing them on a single language number would pick whichever one their question was not about.
4Sources
Counted from the markup of capcut.com and d-id.com, with the spoken claim quoted from the D-ID avatars page. Both rows are kept in the register under one rule, set out on how counted; the column is described on site language.
5Both of them against the rest of the column
| Tool | Site languages |
|---|---|
| CapCut | 27 |
| D-ID | 2 |
| Fliki | None declared |
| Hailuo | None declared |
| HeyGen | None declared |
| Higgsfield | 5 |
| invideo AI | None declared |
| Kling AI | 20 |
| LTX Studio | None declared |
| Pika | None declared |
| PixVerse | 13 |
| Rask AI | One same-host, 15 on subdomains |
| Runway | 5 |
| SceneMixer | 15 |
| Synthesia | 9 |
| VEED | None declared |
| Vidnoz | One same-host, 11 on subdomains |
| Vidu | 9 |
Inclusion rule. Tools with a public English site selling AI video generation or AI video dubbing. Site languages count hreflang alternates served from the host being read; x-default and alternates pointing at another host are not counted, for every vendor alike. Order. Alphabetical by tool name.
Other pairings that turn on the same kind of difference: HeyGen and Rask AI, CapCut and Fliki. The full set of pairings is here; single cells are indexed at entries.
- Site languages27 locales: ar-EG, cs-CZ, de-DE, el-GR, en, es-ES, fi-FI, fr-FR, he-IL, hu-HU, id-ID, it-IT, ja-JP, ko-KR, ms-MY, nl-NL, pl-PL, pt-BR, ro-RO, ru-BY, sv-SE, th-TH, tl-PH, tr-TR, uk-UA, vi-VN, zh-TWdeclared as same-domain alternates
- Site languagesEvery alternate but English is a language-region pair; Russian is declared only as ru-BY and Chinese only as zh-TWhow the codes are written
- Site languages2 languages: de, endeclared as same-domain alternates
- Voice languages“Avatars can narrate or converse in more than 120 languages and accents, with natural pronunciation and expressive delivery”count published without a list