One video maker, ten languages

The ten languages at a glance

Every voice is a natural neural voice with exact timing for every word, so captions and drawings land on cue.

What's the same in every language

• The AI writer. Type one sentence in any language and the script comes back in the language you chose. Scripts in languages other than English are drafted in English first, for accuracy, then translated, so facts don't drift between versions.
• Word-by-word captions. The spoken word is highlighted as it's said, in the right script and direction.
• Drawings chosen by meaning. A multilingual AI model reads each sentence and picks a line drawing from about 10,000 open-source drawings. Because it works on meaning rather than keywords, an Urdu sentence and its Spanish translation can lead to the same picture.
• Motion graphics. Counters for big numbers, charts for growth or decline, timelines and clocks for years, two-sided comparisons and powers-of-ten zooms. Write numbers as plain digits (1500, 86) so they can count up on screen.
• All writer modes. Explainer, lesson with quiz, data story, product explainer, property listing, news recap and how-to.
• Formats. Vertical 9:16, 16:9 YouTube and square 1:1, plus a thumbnail and a carousel export. No watermark.

What's different, and why it matters

Right-to-left languages. Urdu and Arabic captions are laid out right to left, and the highlight moves with the reader's eye. English words inside an Urdu sentence keep their correct order, a common failure in video editors.

Urdu in Nastaliq. Most tools display Urdu in Naskh, the Arabic style, because they don't load a Nastaliq font. Urdu here is in proper Nastaliq, with taller line spacing so the letters aren't clipped. Years and decimals are also read the way Urdu speakers say them: 1929 becomes انیس سو انتیس.

Hindi in Devanagari. Captions are in Devanagari, and you can type your idea in Hindi, Hinglish or English.

Brazilian Portuguese. The Portuguese voice has a Brazilian accent, which matters if your audience is in Portugal or Africa.

Maps. A globe flies to places named in the script. It recognises place names most reliably in English and Urdu, and in the other languages when a name is spelled the same way as in English (Paris, Berlin, Lima, Istanbul, بغداد). When a name isn't recognised, the sentence gets a drawing instead, never a wrong place.

Script length. The same idea takes a different number of words in each language. German and Arabic words carry more meaning each, so a minute of narration has fewer of them; Spanish, Portuguese and Italian need more syllables than English. Each language page has a rough starting point, and the preview always shows the real length.

Making the same video in several languages

Many channels publish one topic in two or three languages. A simple workflow:

1. Make the video in your main language and get the script right.
2. Start a new video with the same idea in the second language. The AI drafts in English and translates, so the facts line up.
3. Check names and numbers in each version, then download each format you need.

Captions for your own clips

Caption my video transcribes your own talking clip (up to 3 minutes, MP4, MOV or WebM, up to 400 MB) with Whisper in any of the ten languages, with every word editable. See Urdu captions, Hindi captions and Arabic captions.

How it works

1. Choose a languagePick one of the ten languages in the studio.
2. Type an idea or paste a scriptWrite one sentence in any language, or paste your own script in the language you chose.
3. Preview and editWatch the video with voice, captions and drawings, and change anything: script, voice, music, theme, speed, format or a scene's drawing.
4. DownloadSign in with Google (free) and download the MP4, with no watermark.

Questions

Updated 2026-11-01