URL Inspection tool: check rendered HTML and resources
Use Search Console's URL Inspection tool to check rendered HTML, the screenshot and page resources, and to read blocked-file and not-indexed results.
By Crawlito · Updated 9 min read
In short
URL Inspection shows Google's view of one URL in a property you own: the HTML, HTTP headers, JavaScript console output and page resources, plus a screenshot in a live test.
The indexed view shows the last crawl and may differ from your live page; the live test fetches the URL now, can show a screenshot, and only confirms whether Google-InspectionTool can access the page.
Google says the screenshot can differ from your page when resources are blocked to Google-InspectionTool, and that it will not render JavaScript from files that robots.txt blocks.
A good result here proves Google can read the page; it says nothing about AI crawlers, which in Vercel's December 2024 analysis did not run JavaScript.
Fetch the page with curl as well: any text that shows up in the tool's HTML but not in curl's output depends on JavaScript.
The URL Inspection tool in Google Search Console shows how Google fetched one page: the HTML it received, the HTTP headers, JavaScript console output, the resources it loaded and, in a live test, a screenshot. Use it to confirm that your main text, canonical tag, robots tags and structured data are present when Google looks at the page. It only answers for Google. It cannot tell you what ChatGPT, Claude or Perplexity read.
This is the step-by-step workflow, with the limits of each view, so you do not trust a green result more than it deserves.
What does URL Inspection show, and what does it not?
You enter a fully qualified URL in the inspection bar of a Search Console property you own, or click an Inspect link in a report. According to Google's URL Inspection help page, the tool gives you information about Google's indexed version of a page and lets you test whether a URL might be indexable. For a property you do not own, Google points you to the Rich Results Test instead.
What you get:
Page indexing status: whether the URL is on Google, with fields such as last crawl, "Crawled as" (desktop or mobile), "Crawl allowed?" (robots.txt), "Page fetch" and "Indexing allowed?" (noindex).
Canonical: the user-declared canonical and the Google-selected canonical.
View crawled page (indexed version): the HTTP request and response and the returned HTML, with page resources and JavaScript console output in the extra response data.
View tested page (live test): a screenshot of the rendered page, the raw HTML returned, the HTTP headers, JavaScript console output and the page resources loaded.
Enhancements: structured data and other features detected at the last index.
What it does not give you: a ranking forecast, a guarantee of appearing in Search results (Google says "URL is on Google" is not one) or any view of what a crawler other than Google's received. Each property also has a daily limit on inspections, and a separate daily limit on live tests.
Indexed version or live test: which should you use?
The tool has two views, and you can toggle between them.
Google index
Live test
What it reflects
The most recently indexed version, which may differ from your live page
A fetch of the URL right now
Screenshot
Not available
Available when the fetch succeeds
Button for the HTML
View crawled page
View tested page
Good for
Finding out what Google actually has
Checking a fix before Google recrawls
Use the indexed view to answer "what does Google have on file?" If the HTML there lacks content that your live page has, Google has not seen the newer version yet. The indexed view shows response data only for URLs with the status "URL is on Google" or "URL is on Google, but has issues", so for a page such as one that is crawled but not indexed, use the live test. Use the live test after a change to answer "can Google read this now?" Google says the live test only confirms whether Google-InspectionTool can access your page for indexing, and it does not test for all possible indexing issues. It does not check sitemaps, referring pages, duplicates or manual actions, and it follows redirects without showing you the final URL. The page must be reachable without a login.
After a fix, the order is: run the live test, confirm the change, then click Request indexing. Google says that submission queues the URL and does not guarantee inclusion, and it usually takes about a day, sometimes longer.
How do you open the crawled HTML and compare it with your page?
Compare three versions of the same page: what your server sends, what Google's tool shows, and what you see in a browser.
Pick a sentence that matters, such as a product description or a price, and copy a short phrase from it.
A count of 0 means your server does not send that text. See raw HTML vs rendered HTML for more ways to see both versions.
Open the page in URL Inspection. Run the live test, click View tested page, and open the HTML it shows. Search it (Ctrl+F or Cmd+F) for the same phrase.
Read the result. If the phrase is missing from curl but present in the tool's HTML, your content depends on JavaScript and Google ran it. If it is missing from both, Google did not get the content either.
One wording point. Google's help page describes the HTML in this panel as the "raw HTML returned", while its JavaScript documentation tells you to use the same tool to check that the rendered HTML has all the content you expect. Rather than argue about labels, rely on the test in step 4: text that appears here but not in curl's output got there through rendering.
Google's JavaScript SEO basics explains why the gap exists. Google queues pages for rendering, and a headless Chromium runs the page's JavaScript when resources allow. Google then parses the rendered HTML for links and uses it to index the page.
What does the screenshot prove, and what does it not?
The screenshot appears only in a successful live test. It is not available for the indexed URL. Google describes it as a screenshot of the rendered page.
It proves that, at the time of the test, Google's fetcher could load and paint the page. A blank or half-built screenshot is a sign to look for a rendering problem.
It does not prove that the content is in the HTML you care about. A screenshot can show text that is drawn in an image or a canvas, which is not text in the page's HTML. It also does not prove the indexed version looks the same, because it comes from a live fetch. And Google states that differences between the screenshot and your page can result from resources that are blocked to Google-InspectionTool, the user agent behind the tool's fetches. Treat the screenshot as a smoke test and read the HTML to confirm.
How do you read the page resources list for blocked or failed files?
The "More info" section lists the page resources Google loaded and the JavaScript console output, per Google's help page. Check it when the screenshot looks wrong or the HTML is missing content.
What to look for:
Scripts and stylesheets that did not load. A missing bundle can leave the page empty or unstyled.
API calls your page makes to fetch content. If the endpoint is blocked or returns an error, the content never arrives.
JavaScript console errors. An exception early in the script can stop everything after it.
Google's help page does not document each status label for a resource, so read what the tool shows for your page rather than relying on a list from memory. The cause that Google does document is robots.txt: Googlebot checks robots.txt before it requests a URL, and Google says it will not render JavaScript from blocked files or on blocked pages. If a script, an API path or a whole folder is disallowed, rendering can stop short.
The fix is to allow Google to fetch the files the page needs to render its main content. Test your robots.txt against each crawler shows how to trace which rule matches a given path. A 403 or 5xx on a resource is a server or firewall problem instead; Google's page indexing help says Googlebot never sends credentials, so a 403 is likely a server setup error.
How do you check canonical, robots tags and structured data?
Do these checks in the HTML panel and the page indexing section.
Canonical. The page indexing section shows the user-declared canonical and the Google-selected canonical. Google notes that the Google-selected value can lag by a few hours. If the two differ, Google chose another URL as the canonical. Google's JavaScript guidance says JavaScript can set a canonical, but you should not use it to change the canonical to a different URL than the one in the original HTML, and you should have only one rel="canonical" tag. Search the HTML panel for rel="canonical" and confirm exactly one match with the URL you intend. More on conflicts: canonical and noindex with JavaScript.
Robots tags. Search the HTML panel for name="robots", and read the HTTP headers the tool shows for X-Robots-Tag. "Indexing allowed?" reports whether a noindex was found. If robots.txt blocks the page, Google cannot see a noindex on it, so "Indexing allowed?" always shows "Yes" in that case. Google also says that when it meets noindex it may skip rendering, so removing a noindex with JavaScript may not work. Take it out of the server HTML. Leftovers from a staging site are common, see staging noindex leaks.
Structured data. Search the HTML panel for application/ld+json to confirm the block is present after rendering; Google says JSON-LD generated by JavaScript can be read. The Enhancements section reports the rich result items detected at the last index, with errors and warnings. It shows only some feature types: Google notes that unsupported rich result types can be valid and appear in Search without showing in this tool.
What do the "not indexed" results mean?
An inspection can end with "URL is not on Google". Expand the status for the reason. Google's page indexing report help defines the common ones:
Status
What Google says it means
Crawled - currently not indexed
Google crawled the page but did not index it
Discovered - currently not indexed
Google knows the URL but has not crawled it yet
Blocked by robots.txt
Your robots.txt blocks the page; this does not guarantee it stays out of the index
Excluded by noindex tag
Google found a noindex directive and did not index the page
Duplicate, Google chose different canonical than user
You declared this page canonical; Google indexed another URL
Soft 404
The page shows a "not found" message but returns a success code
The first two have no tidy fix: Google says it does not index every page, and that "Crawled - currently not indexed" does not necessarily indicate a problem. For a page that matters, inspect it and confirm the main text is in the HTML Google received. The rest map to a concrete change in robots.txt, a tag or a status code.
Why does Google's rendered view not predict what AI crawlers see?
Because the tool shows one crawler's behavior. Google runs a rendering service. In Vercel and MERJ's December 2024 analysis, the major AI crawlers it measured, including OpenAI's, Anthropic's and Perplexity's, did not render JavaScript. Two it named as exceptions were Gemini, which uses Googlebot's infrastructure, and Applebot. That is a snapshot, not a permanent rule, and operators can change it.
So a page can pass URL Inspection and still hand an AI crawler an empty shell. Google's own JavaScript guidance notes that not all bots can run JavaScript and that server-side rendering or pre-rendering is still a good idea. Treat the curl result from step 2 above as the test that applies to crawlers that read only server HTML, and run it for AI user agents as shown in see your page as an AI crawler with curl. For the evidence by crawler, read do AI crawlers run JavaScript?, and for fixes see client-side rendering and the empty HTML problem.
Frequently asked questions
Does URL Inspection show the rendered HTML?
Google's help page calls the panel contents the "raw HTML returned", while its JavaScript documentation says to use the tool to check rendered HTML. To settle it for your page, compare the panel with a curl fetch: text that is only in the panel was added by JavaScript.
Why is there no screenshot?
Google says the screenshot is available only in a live test with a successful result. It is not available for the indexed URL.
Can I use URL Inspection on a site I do not own?
Not for a property you do not have in Search Console. Google recommends the Rich Results Test or the AMP test for URLs you do not own.
Does passing URL Inspection mean AI search tools can read my page?
No. It shows what Google got. Many AI crawlers measured in 2024 did not run JavaScript, so confirm that your key text is in the raw server HTML.
Check what a crawler reads without JavaScript
The free check reads one page the way an AI crawler does, with no account: the robots.txt rules for named AI crawlers, the robots tags and whether your main text is in the raw HTML. For a whole site, an audit runs the checks across up to 500 pages and gives you evidence and prioritized fixes.
Filed under Rendering and content. Topics: · google, rendering, javascript, troubleshooting, search-console
Google can read canonical and robots tags that JavaScript adds, but its docs warn against relying on it. See where raw and rendered HTML conflict, and how to fix it.
Raw HTML vs rendered HTML: what each one is, how to compare them with view-source, curl, DevTools and Search Console, and when the gap hides content from crawlers.
Do AI crawlers render JavaScript? Googlebot does. Most AI crawlers did not in Vercel's 2024 study. What that means for your pages and how to test them.