One page, two spoken-language counts, forty apart
Vidnoz states on one page both that a talking photo can be created in 100+ languages and that it will speak emotively and naturally in 140+ languages. Forty languages separate the two figures and nothing on the page distinguishes them. As of 2026-09-22.
| Figure | The sentence it sits in | A reading the page allows |
|---|---|---|
| 100+ languages | Creating a talking photo | Languages only, with no accents counted |
| 140+ languages | Speaking emotively and naturally | Languages and accents counted together |
| Neither | No list anywhere on the page | Nothing can be checked against a market |
| The larger figure | Repeated in a heading | The one a visitor is most likely to remember |
Inclusion rule. Both published figures for one vendor, with a possible reading recorded as possible rather than asserted. The vendor states neither reading. Order. Smaller figure first, then larger, then what both lack.
1Two numbers, no stated difference
There are benign explanations. One figure might count languages and the other languages plus accents; one might belong to a free tier and the other to a paid one. The page says none of that, so the register does not choose.
Both go in the cell. Printing one would mean picking a number the vendor has not picked, and printing the higher one would mean adopting the marketing figure, which is the habit this register was built to avoid.
2What a visitor actually takes away
The larger figure is repeated in a heading, which is where a reader looks first. In practice the page communicates 140 and mentions 100, and a buyer comparing vendors would carry away the larger one.
That is why a register of this kind records where a figure sits as well as what it says. A number in a heading and a number in body text have different weight, even when both are on the same page.
3Neither figure can be checked against a market
No set is published in either place. A reader who needs to know whether Turkish is among the 140, or the 100, or only one of them, has no page to open and no way to find out without contacting the vendor.
The same vendor declares twelve site alternates, eleven of them on subdomains. So its row needs two numbers in the site cell and two in this one, and not one of the four can be searched for a language.
4Sources
Both figures quoted from the Vidnoz talking photo page. The column itself is described on dialogue language, and every cell Vidnoz fills is on its tool page.
5The same column, tool by tool
| Tool | Voice languages |
|---|---|
| CapCut | No list published |
| D-ID | 120+, count only |
| Fliki | 91 named, plus 117 dialects |
| Hailuo | No list published |
| HeyGen | Two named lists, no total |
| Higgsfield | No list published |
| invideo AI | No list published |
| Kling AI | Five, count only |
| LTX Studio | No list published |
| Pika | No list published |
| PixVerse | No list published |
| Rask AI | 135+, count only |
| Runway | No list published |
| SceneMixer | 15, plus Cantonese for dialogue |
| Synthesia | 143 named |
| VEED | 29 dub-to, 72 detectable |
| Vidnoz | 100+ and 140+, counts only |
| Vidu | No list published |
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.
Neighbouring cells: VEED: dialogue language and SceneMixer: document language. All of them together: the cell index and the register table.
- Voice languagesOne page carries both “in 100+ languages” and “speak emotively and naturally in 140+ languages”two counts on one page
- Site languages12 alternates declared: ar, de, en, es, fr, it, ja, ko, nl, pt, tr, zh-Hantraw count before the rule is applied