Image, Video, and News Sitemaps: Dedicated Channels for Rich Media
A base sitemap carries very little information: here is a URL, and I would like it indexed. It says nothing about what the page contains. When a page embeds images or video, or is itself a news item, what the page contains with regard to displayable media is exactly what decides which presentation slots it can enter. Image packs, video carousels, and news results all form around the media itself.
Media extensions exist for that. They add media-level fields inside the same sitemap: image addresses and captions, thumbnails and player pages for video, headlines and publication dates for news. With those in place a crawler does not have to render the page and guess what is inside; the pairing between media and page is declared directly.
This article covers only the extension layer. Base sitemap structure, truthful lastmod, and URL filtering are already covered in the XML sitemap guide, so one sentence suffices here: the base rules still apply, and an extension adds fields on top of them rather than replacing them. What follows takes the three extensions one by one, then the shared rules, verification, and common mistakes.
1. What each of the three extensions solves
1.1 How base sitemaps and media extensions divide the work
The division fits in one line: the base sitemap says which pages exist; the media extension says which media those pages contain.
The division has practical consequences. An image inlined in an article can already be discovered through HTML, so the extension's value is adding the semantics of which page it belongs to and what it depicts. On a dedicated image page or gallery page, the images are the page, and the extension's job is pairing each image address explicitly with the page that hosts it.
For video the effect is clearer. A video page URL in a base sitemap only says there is a page; the video extension is what tells the crawler where the thumbnail is, where the player page is, and how long the video runs. Presentation needs that information, and HTML does not always expose it reliably.
1.2 Preconditions for each extension
All three carry preconditions, and adding fields without meeting them achieves nothing.
- Image extension: images must be publicly accessible and return 200, and the hosting page must itself be indexable.
- Video extension: the video has a stable player page or content file address, a usable thumbnail exists, and the page carries text rather than only a player.
- News extension: the content is news, the site publishes in that field consistently, and the articles are recent.
The last one is missed most often. The news extension is not a general device for getting new articles indexed faster, and chapter four addresses that specifically.
2. Image sitemaps
2.1 When you actually need one
The value of the image extension is inversely proportional to how easily images can be discovered through HTML. The test is simple.
Cases that need it: the site has dedicated image pages or gallery pages where images are the primary content; the image library is large and spread across the depths of a content management system; images are injected by JavaScript and their addresses are absent from the HTML source; each image has several size variants and the primary one needs to be identified.
Cases that do not: images are inline illustrations, the count is modest, and <img> tags are plainly visible in the source. Adding an image extension there is mostly duplicated effort with near-zero return.
2.2 Syntax and the fields people forget
The image extension declares its namespace on the sitemap root element and then attaches image information inside each URL entry. The core field is the image address; the rest are optional but useful descriptive fields.
<url>
<loc>https://example.com/gallery/edge-cache</loc>
<image:image>
<image:loc>https://example.com/img/edge-cache-01.jpg</image:loc>
<image:title>Request path for an edge cache hit versus an origin fetch</image:title>
<image:caption>A user request served from the edge cache never reaches the origin.</image:caption>
</image:image>
</url>
Two omissions dominate. First, the image address must be an absolute URL and must match what the page actually references. If the page transforms sizes through parameters while the sitemap lists the original file, the two disagree and create ambiguity. Second, the descriptive fields are worth writing: image:title and image:caption are among the few places that state an image's meaning directly, which helps relevance judgements in image search. Compression, format, and loading optimisation belong to another layer, covered in the image optimisation and CDN guide.
3. Video sitemaps
3.1 Required fields
The video extension has more fields than the image one, because video needs more information to be shown. The required items in the official documentation group into four: a thumbnail address, a title, a description, and either a player page address or a content file address. Beyond those, duration and publication date are strongly recommended. For exact field names and validation rules, follow the official documentation.
<url>
<loc>https://example.com/guide/video-seo-basics</loc>
<video:video>
<video:thumbnail_loc>https://example.com/img/video-seo-basics.jpg</video:thumbnail_loc>
<video:title>Minimum configuration for a video page</video:title>
<video:description>Page structure, thumbnail, and player address in the smallest workable setup.</video:description>
<video:player_loc>https://example.com/watch/video-seo-basics</video:player_loc>
<video:duration>412</video:duration>
</video:video>
</url>
Three fields get written wrongly most often. The thumbnail must be a directly accessible image address, not one that requires a login or only appears inside a recommendation widget. The player page address is the page where the video can be watched or embedded, not the video file itself and not the thumbnail. The duration must match the real runtime; a wrong value makes the presentation disagree with the content.
3.2 The split with VideoObject: discovery and eligibility
Getting a video into results takes two gates, handled by different mechanisms.
Discovery comes from the sitemap and page links: they tell the crawler a video exists and where it lives. Eligibility comes from VideoObject structured data on the page: it declares what the video is, its thumbnail, duration, upload date, and whether it can be embedded. Do only the first and the video may be discovered without being eligible for display. Do only the second and the structured data may sit unprocessed for lack of a stable discovery path.
The right approach is therefore both, with consistent information in both places. When the thumbnail address, duration, or title disagree between sitemap and structured data, presentation becomes unstable and diagnosis becomes hard. Field requirements and validation for structured data are in the structured data and rich results guide, how to organise the video content itself is in video content SEO, and the display mechanics of carousels are in the SERP features guide.
4. News sitemaps
4.1 Scope and time constraints
The news sitemap is the most tightly constrained of the three. It applies only to news content, meaning the site is a recognised news publisher producing and updating content in news form. For the overwhelming majority of corporate sites, tool sites, and shops, it does not apply.
It also carries constraints on recency and volume: generally only articles published within a recent window should be included, a single file has a cap on entries, and anything beyond that needs splitting or an index. The specific day window and entry cap follow the current official documentation, and those rules change over time, so check the current version before publishing rather than copying numbers from an older article.
4.2 Why it gets misused more than anything else
The misuse comes from a natural association: a news sitemap looks like a channel that gets new content discovered faster, so people treat it as a general accelerator and put every newly published page inside.
That fails in two ways. First, it violates the scope, so the entries earn no corresponding presentation and the file loses its meaning. Second, it sets wrong expectations: you believe an accelerator is configured when in fact you have configured a list that is not trusted for that purpose, while the internal linking, structure, and content quality work that would actually help gets skipped.
The right stance is that a news sitemap is a dedicated channel for news publishers, not an indexation accelerator. If the site is not a news publisher, what needs fixing is discovery and quality, as described in the website indexation guide.
5. Shared rules
5.1 loc, lastmod, and the two limits
All three extensions share the same underlying rules, and those do not change with the extension.
<loc> must be an absolute URL including scheme and hostname, and it must be the final address rather than one that redirects. The same holds for media addresses inside an extension.
<lastmod> must be truthful. It means the time of the page's last substantive change. Note that the video extension has its own publication date field, and the two must not be conflated: lastmod describes the page while the publication date describes the video.
The two limits are protocol facts: a single sitemap file may contain at most 50,000 URLs and must not exceed 50MB uncompressed. Reaching either limit requires splitting and summarising with a sitemap index. Files carrying media extensions are far larger than plain URL files, so the 50MB ceiling usually arrives first, and estimating by video count is more accurate than estimating by URL count.
5.2 Every listed URL must be accessible and indexable
This looks like the most basic rule and is broken most often: URLs and media addresses listed in a sitemap must return 200 and must be allowed to be indexed.
Three violations dominate. The media file itself carries noindex or requires authentication, for example sitting in private storage or being blocked by hotlink protection. The media address in the sitemap returns a 30x, typically because image processing or a storage migration changed the address while the sitemap stayed stale. The page is indexable while the media directory is blocked in robots.txt. All three nullify the extension, and none of them necessarily shows up in the sitemap report, so they need separate checking.
If media is delivered through a CDN, also consider whether the edge layer changes how those addresses behave. That is the same class of problem as "what the crawler receives may not be your article" discussed in the CDN and SEO guide.
6. Verification
6.1 Sitemap report and media index reports
Verification has two stages: first confirm the file is read correctly, then confirm the media is processed correctly.
Stage one is the sitemap report: whether the file is readable, whether parsing succeeded, how many URLs were processed, and whether field-level errors exist. Extension field errors usually surface here, for example a missing required field or a malformed address.
Stage two is the media index reports: how many images and videos are indexed and how many are excluded for quality or technical reasons. Reading these follows the same logic as handling "discovered but not indexed", described in the Search Console guide.
6.2 Testing a single page for rich media
Reports give an overall state; single-page problems need single-page tests. Pick a few important pages and run them through the rich results testing tool: whether the structured data is recognised, whether required fields are present, whether the thumbnail is reachable, and whether the duration matches.
The value of this step is separating "correct at the sitemap layer" from "correct at the page layer". Very often only one of the two holds: the sitemap is immaculate while the page has no structured data at all, or the structured data is complete while the sitemap still lists an expired address. Verification is finished only when both have been checked.
7. Five common mistakes and a checklist
7.1 Five common mistakes
- Using the thumbnail address as the player page address. The fields mean different things; getting it wrong breaks or distorts the presentation.
- Leaving media files under
noindexor behind authentication. Declaring media you will not serve reduces the extension to nothing. - Letting listed media addresses return 30x. Usually the address changed after image processing or a storage migration while the sitemap did not follow.
- Publishing a duration that does not match the video. The presentation disagrees with the content; at best it is corrected, at worst eligibility is lost.
- Treating the news sitemap as a general accelerator. Outside its scope, no configuration produces the corresponding presentation.
7.2 Checklist
- Base sitemap rules still hold: no 404s, no
noindexpages, no blocked URLs, truthfullastmod. - Pages that need an extension are identified: dedicated image pages, gallery pages, video pages, news pages.
- Image and media addresses are absolute and match what the pages actually reference.
- The four required video field groups are present, and duration and publication date match reality.
- Sitemap and
VideoObjectagree, with no contradicting fields between them. - The news extension is used only for news content, with the recency window and entry cap following current official documentation.
- No single file exceeds 50,000 entries or 50MB, and anything beyond is summarised by an index.
- Every listed media address returns 200, is publicly reachable, and is not blocked by robots rules.
- The sitemap report shows no parsing errors, and media index reports have a baseline with periodic reviews.
7.3 Deployment order
Do "can be accessed" before "can be understood". In other words, confirm every media address returns 200, is publicly reachable, and is unblocked, and only then fill in the fields. Reverse that order and you reach an awkward place: the fields are complete, the report is full of errors, and the cause turns out to be unreachable addresses.
The second priority is the dual configuration for video: sitemap and structured data both in place and consistent. It is the only one of the three extensions that needs two locations to cooperate, and the easiest place to end up half right. The image extension can wait, because its return is inversely proportional to how naturally images can be discovered, and for many sites basic checks are enough. Before going live, run the crawl and index items in the technical SEO checklist alongside this, which blocks most of the basic errors.
Source: 16IDC original. Crawling and indexing behaviour, SERP presentation mechanics, and the sitemap protocol follow the official Google Search Central documentation and the sitemaps.org protocol.
Code Examples
Example 1: Declaring image and video namespaces with entries
<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
xmlns:image="http://www.google.com/schemas/sitemap-image/1.1"
xmlns:video="http://www.google.com/schemas/sitemap-video/1.1">
<url>
<loc>https://example.com/blog/video-guide</loc>
<image:image>
<image:loc>https://cdn.example.com/img/video-cover.jpg</image:loc>
<image:title>Site structure tour cover</image:title>
<image:caption>Three minutes on directory levels</image:caption>
</image:image>
<video:video>
<video:thumbnail_loc>https://cdn.example.com/img/thumb.jpg</video:thumbnail_loc>
<video:title>Site structure tour</video:title>
<video:description>Directory levels and internal link design.</video:description>
<video:content_loc>https://cdn.example.com/video/guide.mp4</video:content_loc>
<video:duration>180</video:duration>
</video:video>
</url>
</urlset>
The first snippet declares media inside each page entry so crawlers map media to pages without rendering.
Example 2: News entry plus pre-publish validation
# The news extension needs its own namespace declaration:
# xmlns:news="http://www.google.com/schemas/sitemap-news/0.9"
# publication_date must be ISO 8601 and match the visible publish time
# Validate the structure first
xmllint --noout --schema \
https://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd \
sitemap-media.xml
# Then confirm it is reachable in production
curl -sI https://example.com/sitemap-media.xml | grep -Ei '^(HTTP/|content-type)'
# Count entries to catch accidental deletions or duplicates
curl -s https://example.com/sitemap-media.xml | grep -c '<url>'
The second catches format problems with schema validation and a response check before publishing.
Community & Resources
Official docs and communities
- Sitemap protocol — Authoritative definition of the root element, namespaces and base fields.
- Image sitemaps — Required and optional fields for image entries.
- Video sitemaps — Requirements for thumbnails, player pages and duration.
- News sitemaps — Scope, publication dates and eligibility rules.
Learning resources and tools
- Build a sitemap — File size limits, index files and submission.
- Large sitemaps — Splitting by 50000 URLs or 50MB and sharding.
- Bing webmaster guidelines — Another major engine's sitemap expectations.
- sitemap questions — Hands-on questions on namespaces, encoding and validator errors.
Original source: 16IDC ↗