# Meta name googlebot-image: why Google ignores it
> Google accepts only googlebot and googlebot-news in a robots meta tag, so meta name googlebot-image does nothing, while robots.txt accepts it.
Canonical: https://brandonlazovic.dev/articles/googlebot-image-meta-tag-ignored/
Author: Brandon Lazovic
Published: 2026-09-29
## The short version
- Google ignores a robots meta tag named googlebot-image, because the tag accepts only two crawler names, googlebot and googlebot-news.
- The same name works as a robots.txt group, and a googlebot group there also reaches Googlebot-Image, which answers to both names.
- An X-Robots-Tag header can carry a crawler name before its rule, but Google publishes no list of the names the header accepts.
- This site's request log shows 10 Google names in 68 days, and a robots meta tag can address only one of them by name.
A robots meta tag is a line in a page's HTML that tells search engines what they may do with that page, such as leave it out of their results. A crawler, the program a search engine sends to fetch pages, answers to a short name in such rules, which Google calls a user agent token. Google ignores the tag when its `name` is `googlebot-image`, the token of the crawler that fetches pictures for Google Images. Its documentation says Google "supports two user agent tokens in the robots meta tag; other values are ignored," and the two are `googlebot` and `googlebot-news`. [1] The same name works in robots.txt, the file at a site's root that tells crawlers which paths they may fetch, and there Google's image crawler follows it. [2] So a rule aimed at the image crawler does nothing in one place and works in another, and the ignored version looks exactly like one that works. If you wrote that tag to keep pictures out of Google Images, it never did.
> Google accepts two crawler names in a robots meta tag. This site's request log shows 10 Google names in 68 days, and the tag can address only one of them by name.
## What does Google do with meta name="googlebot-image"?
Google ignores a robots meta tag whose name is `googlebot-image`. Google's robots meta tag documentation lists two crawler names the tag accepts, `googlebot` for text results and `googlebot-news` for news results, and says other values are ignored. [1] A tag written as `` therefore changes nothing in Google, while the same tag named `googlebot` or `robots` does take effect.
Putting one crawler's token in the `name` attribute is meant to narrow a rule to that crawler, and in this tag it works for only two tokens. The attribute usually reads `robots`, and Google says that value "specifies that the rule applies to all crawlers." [1]
A token is the short label a rule uses. It is a different thing from the longer user agent string a crawler sends with every request, which is what the request log later in this piece counts.
My read is that the two accepted names describe kinds of results, and not crawlers. Google glosses each one as a result type, "for all text results" and "for news results." [1] Googlebot-News does not even send requests under its own name. Google says it "doesn't have a separate HTTP request user agent string" and crawls with Googlebot's instead. [2] An image crawler has no slot in a list organized that way.
## Which crawler names does a robots.txt file accept?
A robots.txt file accepts every crawler name Google lists, and one crawler can answer to more than one name. The file sits at a site's root and tells crawlers which paths they may fetch, in groups that each start with a `User-agent` line. Google lists Googlebot-Image under two tokens, `Googlebot-Image` and `Googlebot`, so a group naming `googlebot` also governs Google's image crawling. [2]
Google states the rule in one sentence, "Some crawlers have more than one user agent token; you need to match only one crawler token for a rule to apply." [2]
Google-InspectionTool, which runs the URL Inspection tool in Search Console and the Rich Results Test, lists `Googlebot` as a second token. So does Google-CloudVertexBot, which crawls when site owners ask it to while building agents on Google Cloud's Vertex AI. GoogleOther, a general-purpose crawler Google uses for research fetches, lists only its own name. [2]
A crawler follows exactly one group, the one "with the most specific user agent that matches the crawler's user agent." [3] Where a site has both a `googlebot` group and a `googlebot-image` group, the image crawler takes the second and ignores the first.
The fallback follows Google's token list, never a name that looks similar. In Google's own example, Storebot-Google follows the wildcard group, written `User-agent: *` and meant for any crawler without a group of its own, "because there is no specific Storebot-Google group," even though a `googlebot` group sits in the same file. [3]
Three kinds of Google requester sit outside robots.txt groups altogether.
- AdsBot-Google, which checks web page ad quality for Google Ads, skips the wildcard group. In Google's words, "The global user agent (*) is ignored." [4]
- Google-Safety, which looks for malware, "ignores robots.txt rules." [4]
- Fetchers a person triggers, such as Google Read Aloud, "generally ignore robots.txt rules" because a user asked for the page. [5]
This site's robots.txt holds a single group, `User-agent: *`, allowing everything. By Google's list, that group never addresses the two AdsBot names that visited. Here it costs nothing, since every path is open anyway. On a site that disallows a path in the wildcard group to keep all crawlers out, AdsBot keeps fetching it until a group names it. How allow and disallow lines resolve inside one group is covered in [robots.txt allow versus disallow](/articles/robots-txt-allow-vs-disallow/).
## Can an X-Robots-Tag header name one crawler?
An X-Robots-Tag header can name one crawler by putting that crawler's name before the rule. The header is a line the server sends alongside a file, carrying the same rules a robots meta tag carries. Google documents that it "may optionally specify a user agent before the rules," with `googlebot: nofollow` as the example, and that "Rules specified without a user agent are valid for all crawlers." [1]
The header matters most for files that have no HTML to hold a meta tag. Google's instruction is direct: "To block indexing of non-HTML resources, such as PDF files, video files, or image files, use the X-Robots-Tag response header instead." [1]
What Google does not publish is a list of the names the header accepts. The two-name limit is stated for the meta tag. The header section shows `googlebot` and a placeholder called `otherbot`, and stops there. [1] So whether `X-Robots-Tag: googlebot-image: noindex` works is undocumented. I have not tested it either, since a clean test means watching an image's indexing status for weeks.
This site sets its headers in one file, `public/_headers`, where each block of rules is keyed by a path pattern. Four of those blocks send `X-Robots-Tag: noindex` with no crawler name in front. Three cover the plain-text copies of pages kept for AI assistants, and one covers the skill files published for AI agents.
A live request to one of those page copies on September 22, 2026 came back carrying that noindex header, so the rule ships as written. With no name, each rule applies to every crawler, which is the intent. A file keyed by path can still target one crawler, because the crawler's name travels inside the header's value.
## Which Google crawlers actually visit this site?
Ten Google names appeared in this site's request log between July 16 and September 21, 2026. The log is a daily count from Cloudflare, the network that serves this site, of requests per user agent string. That string is the full line of text each visitor sends to describe itself, longer than the short token a rule uses, and each row was counted under one name only.
Googlebot led the count. Every figure is a claim about identity, since nobody here checked these requests against Google's published address ranges.
The table below lists each name, its requests, the robots.txt tokens Google lists for it, and whether a robots meta tag can address it by name. [2] [4] [5]
| Name in the log | Requests | Tokens Google lists for robots.txt | Meta tag can name it |
|---|---|---|---|
| Googlebot | 4,748 | `Googlebot` | Yes, as `googlebot` |
| Googlebot-Image | 811 | `Googlebot-Image`, `Googlebot` | No |
| GoogleOther | 735 | `GoogleOther` | No |
| Google-CloudVertexBot | 477 | `Google-CloudVertexBot`, `Googlebot` | No |
| Google-Extended | 265 | `Google-Extended`, as a control token only | No, and no real request carries it |
| Google-Safety | 57 | None, since it ignores robots.txt | No |
| Google-InspectionTool | 23 | `Google-InspectionTool`, `Googlebot` | No |
| Google-Read-Aloud | 20 | None, since it generally ignores robots.txt | No |
| AdsBot-Google | 6 | `AdsBot-Google`, and it skips the wildcard group | Google suggests it, in a separate passage |
| AdsBot-Google-Mobile | 3 | `AdsBot-Google-Mobile`, and it skips the wildcard group | Not stated |
The AdsBot row carries the one sentence in Google's meta tag page that pulls against its own two-name rule. The same page that says other values are ignored also says "To block non-search crawlers, such as AdsBot-Google, you might need to add rules targeted to the specific crawler," with `` as its example. [1]
Both sentences are Google's. My read, unconfirmed by Google: the two-name limit covers Google Search, and the AdsBot example is an exception for a crawler outside it.
The Google-Extended row is 17 daily rows of impostors. Google says that token "doesn't have a separate HTTP request user agent string," and uses it only as a switch for how Google may use content. [2] So every one of those 265 requests announced a name no Google crawler sends, and 148 of them arrived on September 11 alone. A user agent is a claim, which is why [a user agent claiming to be Googlebot proves nothing by itself](/articles/a-user-agent-claiming-googlebot-proves-nothing/).
The interactive tool at the end of this article takes each of the three places and each of four names, and marks which of these ten names the rule is addressed to, by Google's own rule for that place.
## How do you keep images out of Google Images?
To keep images out of Google Images, name the image crawler in robots.txt or send a noindex header on the image files themselves. A robots.txt group for `Googlebot-Image` that disallows a path stops Google fetching the images there. An `X-Robots-Tag: noindex` header on each image file is Google's documented route for files with no HTML. A meta tag named `googlebot-image` does neither.
Three routes work, each suited to a different case:
1. **A robots.txt group.** Add `User-agent: Googlebot-Image` followed by `Disallow: /private-images/`. Google says preferences addressed to that crawler "affect Google Images, Discover, Google Video, and all features in Google Search where images, logos, and favicons are presented." [2] A broad disallow for Googlebot-Image therefore reaches further than the Images tab, into the logos and favicons Google shows in ordinary search results.
2. **A header on the image file.** Send `X-Robots-Tag: noindex` with the image itself, which works on files that cannot hold a meta tag. [1]
3. **A page-level rule.** Add `noimageindex` to a meta tag named `robots` or `googlebot`. Google defines it as "Do not index images on this page." [1]
Pick one of the first two per file. A crawler blocked by robots.txt never fetches the file, so a header sent with it goes unread, which is the mechanism behind [Disallow stops the crawl, not the index](/articles/robots-disallow-is-not-noindex/).
## Sources
1. Google Search Central: Robots meta tag, data-nosnippet, and X-Robots-Tag specifications (updated Mar 24, 2026) — https://developers.google.com/search/docs/crawling-indexing/robots-meta-tag
2. Google Search Central: Google's common crawlers (updated Jul 14, 2026) — https://developers.google.com/search/docs/crawling-indexing/google-common-crawlers
3. Google Search Central: How Google interprets the robots.txt specification (updated Aug 31, 2026) — https://developers.google.com/search/docs/crawling-indexing/robots/robots_txt
4. Google Search Central: Google special-case crawlers (updated Sep 17, 2026) — https://developers.google.com/search/docs/crawling-indexing/google-special-case-crawlers
5. Google Search Central: Google user-triggered fetchers (updated Aug 19, 2026) — https://developers.google.com/search/docs/crawling-indexing/google-user-triggered-fetchers