By Zach Wennstedt, founder · Updated November 6, 2026
Yes, Google can index a video you host yourself. Give each MP4 its own watch page, mark it up with VideoObject including a contentUrl that points at a stable file Googlebot can fetch, serve a thumbnail Googlebot can reach, list it in a video sitemap, and don’t block the file in robots.txt. YouTube isn’t required. It’s just the path with the fewest moving parts.
This is the companion to our post on VideoObject schema and fast YouTube embeds. That one covers video hosted on YouTube. This one covers the other case: the file lives on your server or CDN, and you want it indexed and fast.
Does Google index self-hosted video?
It does, with conditions. Google’s video documentation says it processes common file types including MP4, WebM, MOV and M3U8, that data URLs aren’t supported, and that the file must be available at a stable URL. It also says Google needs to fetch the actual bytes of the file to enable features like video previews and key moments. So the real question isn’t “can I self-host” but “can Googlebot reach the bytes, on a page that qualifies?”
What is a watch page, and why does it decide everything?
A watch page is a page whose main purpose is to show one video. Google’s documentation says the video must be embedded on one, and that a page where the video is complementary to the rest of the content doesn’t count, giving the example of a blog post that reviews an embedded video. This is the most common reason self-hosted video stays unindexed.
| Page type | Qualifies? | What to do |
|---|---|---|
| One URL, one video, unique title and description | Yes | Mark it up and put it in the video sitemap |
| Service page with an explainer video beside the copy | No, the video is complementary | Keep it for visitors; make a separate watch page if you want it indexed |
| Blog post that discusses a video | No (Google’s own example) | Link to the watch page from the post |
Google also asks that each watch page have a title and description unique to that video. In practice: a thin watch page that holds the player, a heading, a two-sentence summary and the transcript beats a stuffed page where the video is one element of twelve.
Which VideoObject properties do you need?
Google’s structured data documentation requires three: name, thumbnailUrl and uploadDate (ISO 8601, and it recommends including the timezone). For self-hosted files the property that matters most is the recommended one: it says to provide contentUrl if possible, and to use embedUrl only as the alternative when the file URL isn’t available.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "VideoObject",
"name": "How we replace a water heater in Aurora",
"description": "A two-minute walkthrough of a tank replacement, start to finish.",
"thumbnailUrl": "https://example.com/video/water-heater-poster.jpg",
"uploadDate": "2026-11-06T09:00:00-07:00",
"duration": "PT2M10S",
"contentUrl": "https://cdn.example.com/video/water-heater.mp4"
}
</script>
The same documentation recommends duration in ISO 8601 and lists hasPart for marking important segments. Validate it, then check the watch page in Search Console’s URL Inspection.
How should the HTML5 video element be set up?
For a video that visitors choose to play, use preload="none" with a poster. The MDN reference defines none as “should not be preloaded,” notes the default varies by browser, and says the attribute is a hint rather than a command. web.dev’s lazy-loading video guide calls preload="none" the best chance of avoiding the video download.
<video controls playsinline preload="none"
width="1280" height="720"
poster="/video/water-heater-poster.jpg">
<source src="https://cdn.example.com/video/water-heater.mp4" type="video/mp4">
</video>
- Always set width and height. MDN recommends explicit dimensions on every video to avoid layout shift.
- The poster is your LCP element. web.dev notes videos can be LCP candidates and recommends a poster for non-autoplay video. It also says you can’t set
fetchpriorityon theposterattribute itself, so if the video is above the fold, preload the poster image separately with high priority. - Don’t hide the file behind a click. Google’s documentation says not to rely on user actions like clicking to load the video. A facade that injects the player after a click can be fine for visitors, but your watch page should still have the video element (or the structured data pointing at the file) in the initial HTML. If the player only exists after JavaScript, see our notes on JavaScript SEO.
- Make the file seekable fast. Encoding with the metadata at the front of the file (ffmpeg’s
-movflags +faststart) lets playback begin before the whole MP4 arrives.
What does a video sitemap need?
Google lists three ways to give it video metadata: structured data, a video sitemap and Open Graph. A sitemap is the one that scales when you have many watch pages. Per Google’s video sitemap documentation, the required tags are video:thumbnail_loc, video:title and video:description, plus either video:content_loc or video:player_loc.
<url>
<loc>https://example.com/videos/water-heater-replacement/</loc>
<video:video>
<video:thumbnail_loc>https://example.com/video/water-heater-poster.jpg</video:thumbnail_loc>
<video:title>How we replace a water heater in Aurora</video:title>
<video:description>A two-minute walkthrough of a tank replacement.</video:description>
<video:content_loc>https://cdn.example.com/video/water-heater.mp4</video:content_loc>
<video:duration>130</video:duration>
</video:video>
</url>
| Rule from Google’s documentation | Value |
|---|---|
video:duration | 1 to 28,800 seconds (8 hours) |
video:description | Maximum 2,048 characters |
| Tags per video | Maximum 32 |
video:content_loc | Must not equal the page’s <loc> URL; HTML isn’t a supported format |
| Access | Everything referenced must be reachable by Googlebot, not behind a firewall, over HTTP or FTP (streaming protocols aren’t supported) |
Google recommends content_loc over player_loc when you can. The page URL goes in <loc>, which is the watch page, not the MP4. Submit the sitemap in Search Console, then use the Video indexing report, which shows how many indexed watch pages contain an indexed video and why others didn’t.
Can the video live on a CDN or another domain?
Yes, with discipline. Google’s documentation says the video host and streaming server need enough resources to support crawling. If you use a CDN, you might get Search Console notifications about crawl errors, and to receive them you need to verify ownership of the URL space on the CDN, which has to be a subdomain or subdirectory reserved for your content. Plan for these failure modes:
- Blocked bytes. Google says not to block the video file with robots.txt or noindex. Check robots.txt on the CDN hostname too. It’s a separate host with its own file.
- Expiring URLs. “Stable URL” is Google’s wording. A signed link that changes every hour isn’t stable, so give Googlebot a permanent file path and sign only the playback copy if you must.
- Hotlink and bot protection. A rule that returns 403 to unknown user agents will hide the file from Googlebot. Test with a real fetch:
curl -IthecontentUrland confirm a 200 and a video content type. - Thumbnails on the same terms. Google says the thumbnail must be accessible by Googlebot and Googlebot Images, and must be at least 60x30 pixels (larger preferred).
Keep the watch page on your main domain so the ranking signal and the internal links accrue to your site, even if the bytes come from a CDN subdomain. Page weight still matters either way; our site speed work covers how we keep video pages inside Core Web Vitals.
What if the video still isn’t indexed?
Work down this list in order. It goes from the cheapest check to the most involved, and each step maps to something Google’s documentation says.
- Is it a watch page? If the video shares the page with a lot of other content, give it its own URL first.
- Is the page itself indexed? A video can’t be indexed from a page Google hasn’t indexed. Run URL Inspection.
- Does the markup parse? Run the JSON-LD through the Rich Results Test and confirm
uploadDatecarries a timezone. - Can Googlebot fetch both files? Request the thumbnail and the
contentUrlfrom outside your network, and check robots.txt on whichever hostname serves them. - Is the sitemap entry valid?
content_loccan’t equal the page URL, and the description and duration have to stay inside the limits above.
Then wait. Indexing isn’t instant, and the Video indexing report is where the reasons show up.
What is the short checklist?
- One watch page per video, with a title and description unique to that video.
VideoObjectwith name, thumbnailUrl, uploadDate, plus duration andcontentUrl.- Thumbnail and video file both fetchable by Googlebot, no robots.txt block, stable URLs.
<video preload="none" poster width height>, with the poster treated as the LCP element.- Video sitemap submitted, then monitored in the Video indexing report.
- Transcript printed on the page. It helps visitors and gives search engines text to match.
If you’d rather not build the pipeline yourself, Eye To Ad Media offers video SEO as a service, and our technical SEO team can implement the watch pages and sitemap on your stack.
