What search engines need from a translated video
A search engine cannot watch a video the way a viewer does. It relies on the page around the video: the URL, the title, the visible text, the structured data and the links between language versions. A German dub embedded on an English page with an English title gives a German searcher almost nothing to match.
So international video SEO is mostly about the page, not the video file. The goals are simple to state:
- Each language version has a page a search engine can index separately.
- Each page clearly signals its language and its relationship to the other versions.
- The text a searcher would type appears on the page in their language.
- The video itself is easy for crawlers to discover and describe.
This article covers your own website. Titles, descriptions and tags on YouTube follow different rules, covered in translating YouTube titles and descriptions.
One page per language
The foundation is a separate URL for each language version. Common structures:
- Subdirectory
- example.com/de/product-tour, simple to manage on one domain
- Subdomain
- de.example.com/product-tour, sometimes used when teams or hosting differ by market
- Country domain
- example.de/product-tour, a strong regional signal but more domains to maintain
- URL parameter
- example.com/product-tour?lang=de, generally harder to manage and not recommended by most SEO guides
Whichever you choose, avoid serving different languages at the same URL based on the visitor's browser settings or location. Crawlers usually arrive without those signals and may only ever see one version. A visible language switcher that links to the other URLs is better than automatic redirection.
Each page should be genuinely in its language: translated headings, body text, button labels and alt text, not just a translated video with an English page around it.
hreflang annotations
hreflang tells search engines that several pages are versions of the same content in different languages or regions. It can be declared in the HTML head with alternate link elements, in HTTP headers, or in an XML sitemap.
Rules that commonly trip people up:
- Every page in the set must list all versions, including itself. If the German page points to the French page but not the reverse, the annotation may be ignored.
- Use language codes (such as de, ja, pt) and optionally a region (such as pt-BR or es-MX). A region alone is not valid.
- Add an x-default entry for the page visitors should see when no language matches, often a language-selector page or the source-language version.
- Point hreflang only at indexable pages that return a normal response, not at redirects or pages blocked from indexing.
Search engines differ in how much weight they give hreflang, and their guidance changes, so check each engine's current documentation. hreflang helps the right version appear for the right audience; it does not make a page rank where it otherwise would not.
Translated titles, descriptions and on-page copy
Translate metadata for searchers, not word for word. The page title and meta description should use the terms people in that market actually search for, which may differ from a literal translation of your source title. A native-speaking reviewer or local keyword research helps here.
Give each version:
- A translated page title and meta description.
- A translated H1 and introduction that say what the video shows.
- A translated video title and description in the player or structured data.
- Translated alt text on the thumbnail image, and ideally a localized thumbnail if the original has text on it.
Product names and terms should follow your glossary so the page, the video and the transcript use the same words.
Video structured data per language
Schema.org's VideoObject type lets you describe a video in machine-readable form. Google documents properties such as name, description, thumbnailUrl and uploadDate, along with contentUrl or embedUrl, and schema.org also defines inLanguage and transcript. Check the current documentation for which properties each search engine requires or recommends, since lists change.
For translated videos:
- Add a separate VideoObject on each language page, not one block listing every language.
- Write name and description in the page's language.
- Set inLanguage to the spoken language of that version, such as de for a German dub.
- Point contentUrl or embedUrl to that language's video file or player, not the source version.
- If the thumbnail contains text, provide a localized thumbnail URL.
A video sitemap is an optional extra that lists video details for crawlers, useful when videos load in ways crawlers might miss.
Transcripts and captions on the page
A transcript on the page puts every spoken word into indexable text in the page's language. It also helps viewers who prefer reading, need to skim, or cannot play sound. Place it below the video, either in full or in an expandable section, and edit it lightly for readability. The article on publishing transcripts on your website covers layout and editing.
Captions in the player, loaded as a WebVTT caption track, help viewers and accessibility, but search engines may not read caption files attached to a player. Treat the on-page transcript as the text for search and the caption track as the text for viewing. For the YouTube side of captions and discovery, see YouTube captions and SEO.
A software company has a five-minute English product tour on example.com/tour. It creates example.com/de/tour and example.com/fr/tour, each with a dubbed video, a translated page title and introduction, a German or French transcript below the player, and a VTT caption track in the same language. All three pages carry hreflang links to each other and an x-default pointing to the English page. Each has its own VideoObject with inLanguage set to en, de or fr and a localized thumbnail. The team then watches search impressions for each URL in its search console tool rather than expecting any particular ranking.
Hosting choices and their trade-offs
Where the video file lives affects what search engines see.
- Self-hosting on your site or a CDN gives full control over the file, the player and the structured data, but you handle encoding, bandwidth and player compatibility.
- A dedicated video hosting platform usually provides a player, adaptive streaming and analytics, and some let the video be attributed to your page rather than the platform's. Check how each handles separate language versions and caption tracks.
- Embedding from a public video platform is easy and brings that platform's audience, but the platform's own watch page can appear in search alongside or instead of yours.
There is no universally right answer. Teams often publish on a public platform for reach and embed or self-host on their own pages for search and conversion, accepting some overlap. Player and caption-track options are covered in offering several language versions of a video on your site.
Pitfalls and limits
- A translated video on a page whose text is still in the source language.
- hreflang sets that are not reciprocal, or that point to redirected URLs.
- One VideoObject on a page listing a different language from the one actually spoken.
- Automatic redirects based on browser language, which hide versions from crawlers.
- Machine-translated metadata nobody reviewed, using terms no local searcher types.
- Expecting results quickly. Indexing and ranking depend on many factors beyond these signals, and no setup guarantees visibility.
What mydubly provides for multilingual video pages
mydubly produces several of the assets each language page needs. A dubbed job gives a translated video with the original picture and a translated voice track, target-language SRT and VTT subtitles that follow the dubbed lines, and transcripts in both languages. A transcript job gives a plain transcript, a timestamped transcript and subtitles, optionally translated. The VTT file can be loaded as a caption track in your player, and the transcript text can go on the page; the subtitle generator page describes those outputs, and the video localization page covers the full workflow.
It does not write page titles, meta descriptions, hreflang or structured data, and it does not translate on-screen text or thumbnails. Its translated video contains no embedded subtitle track; captions are separate files. Most outputs are MP4, but when the source video codec cannot go into MP4 the output is MKV, which may not play inline in every browser, so check the format before embedding it on a page.
Next step: audit one video
Pick one translated video and check its page: own URL, translated text, reciprocal hreflang, a VideoObject in the right language, and a transcript on the page. Fix what is missing, then apply the same template to every other language version.
Frequently asked questions
Do translated videos need separate pages for each language?
For search purposes, yes. A separate URL per language lets search engines index and show the right version to the right audience. A single page that switches language based on browser settings risks crawlers only ever seeing one version.
Does hreflang improve rankings?
hreflang helps search engines show the appropriate language version to each audience; it is not a ranking boost in itself. It works only when every page in the set lists all the others and itself. Check each search engine's current documentation, as their use of the signal differs.
Should structured data list all language versions on one page?
Usually not. Add a separate VideoObject to each language page, written in that page's language, with inLanguage set to the language spoken in that version and contentUrl or embedUrl pointing to that version's file or player.
Do search engines read subtitle files?
You should not rely on it. Caption tracks attached to a player may not be read by search engines. Publishing the transcript as visible text on the page is the dependable way to make the spoken content indexable, while the caption track serves viewers.
Is it better to self-host translated videos or embed them from a video platform?
It depends on your goals. Self-hosting gives full control over the page and markup; public platforms bring reach but their watch pages may appear in search instead of yours. Many teams do both: publish on a platform for its audience and host or embed on their own pages for search and conversion.
Can machine translation be used for video page metadata?
It can produce a first draft, but have a native speaker review titles and descriptions. Searchers in each market often use different terms from a literal translation of your source title, and reviewed metadata is more likely to match what they type.