The URL Inspection tool is a diagnostic feature inside Google Search Console that reports what Google holds about a single web address: whether the page was crawled, which copy was indexed, which canonical Google selected and, where nothing was stored, the stated reason. It exists because site owners cannot look into Google's index from outside. Aggregate reports describe patterns across thousands of pages; this feature answers the narrower question of what happened to one. Use requires a verified Search Console property, and the address inspected must belong to it.

What the report contains

By default the feature shows the last indexed version of a page, not the live one, according to Google's help documentation. A verdict sits at the top: "URL is on Google", "URL is on Google, but has issues", "URL is not on Google" or "URL is an alternate version". The first means eligibility to appear in results, not appearance; Google advises searching for the address to confirm.

Fields below follow the order in which Google processes a page. Discovery lists sitemaps that name the address, though only those submitted through the Sitemaps report or declared in robots.txt, and a referring page that Google possibly used to find it. Crawl gives the last crawl time, whether a desktop or smartphone crawler visited, whether robots.txt allowed the fetch, the page fetch result and whether a noindex directive permits indexing. Indexing sets the canonical declared by the page against the one Google selected; a mismatch means Google folded the page into another address.

Enhancement sections follow. The help page lists 26 supported rich result types, from breadcrumbs and events to job postings, recipes and vehicle listings, alongside video indexing and linked AMP pages. Choosing View crawled page exposes the raw HTML, HTTP headers, JavaScript console output and loaded resources. That view reflects the fetch-then-render design Google documented in December 2024: the crawler retrieves the parent HTML, then the Web Rendering Service downloads referenced resources and builds the page as a browser would. A screenshot exists only for live tests.

The live test and Request Indexing

Test live URL fetches the page in real time with the Google-InspectionTool user agent. Google's crawler reference says that agent affects only Search testing tools, such as the Rich Results Test and this inspector, and has no effect on Search itself. It is a separate fetcher from Googlebot, the crawler that feeds the index.

The test follows redirects without saying so and cannot judge states decided at indexing time: "crawled - currently not indexed", "discovered - currently not indexed", duplicate and canonical selections, and pages indexed without content. Manual actions, security issues and quality guidelines also fall outside it. A passing result therefore means only that the page can be fetched and parsed. Site-wide faults, among them DNS errors, invalid certificates, an unreachable robots.txt and a "Hostload exceeded" status that signals a server at capacity, can block the test altogether.

Request Indexing appears once the live check passes; a page judged non-indexable cannot be submitted. Google says indexing typically takes a day or so but can take a week or two, guarantees nothing, and applies a daily limit without publishing it. For many pages the help page recommends a sitemap carrying lastmod values instead.

Programmatic access

The URL Inspection API arrived on January 31, 2022, in a Search Central post by Daniel Waisberg and Dori Rosenberg. It returns data about the indexed version only. The response mirrors the interface fields, from verdict and last crawl time to both canonicals, with AMP and rich results blocks; the mobile usability block is marked deprecated. According to Google's usage limits page, last updated on August 28, 2025, each property may make 2,000 queries a day and 600 a minute, while a Cloud project is capped at 10 million a day and 15,000 a minute. A publisher with 1,192 pages, the size in the April 2026 case described below, could therefore inspect every address within a day. The reference documents a single method, index.inspect, and no endpoint for requesting indexing.

Origin and evolution

The lineage runs through Fetch as Google, the legacy Webmaster Tools feature that fetched and rendered pages and could submit them to the index; Google explained its rendering mode in May 2014. The replacement was announced on June 25, 2018, in the beta Search Console, in answer to what Google called one of its most common user requests: more detail on how Search sees a specific URL, according to the launch post. The first version reported the last crawl date and status, errors, the canonical and noindex detection. In January 2019 it added HTTP response, page resources, JavaScript logs and rendered screenshots, according to Search Engine Land. Fetch as Google disappeared on March 28, 2019, according to Bill Hartzer, and in August 2019 Google moved its testing tools to the evergreen Googlebot rendering engine.

Request Indexing went offline on October 14, 2020 and returned 69 days later, according to Search Engine Land. Google restored the button in December 2020, with a reminder that large sets of addresses belong in a sitemap. On October 11, 2021, Google aligned its public testing tools with the inspector's engine. The Mobile-Friendly Test and the Mobile Usability report were retired on December 1, 2023. Microsoft's rebuilt Bing Webmaster Tools added a beta equivalent in July 2020 that, according to Search Engine Land, also compares indexed and live versions.

Why it matters to marketers

Organic landing pages, campaign microsites and content programmes earn traffic only once indexed, and this feature is the single place where Google states, address by address, whether that happened and why not. PPC Land's explainer on the reporting platform describes Search Console as the only official channel between Google Search and the people publishing into it. The inspector also anchors a routine Google recommends for structured data: validate markup in the Rich Results Test, deploy the pages, then inspect them, as PPC Land noted when Google expanded the education Q&A carousel.

Google's public stance adds a reading of its own. John Mueller said on July 15, 2025 that a site barely indexed on sound hosting often signals that Google's systems are unconvinced about it overall.

Limitations and disputes

The indexed view is a snapshot of the last crawl, so it can trail the live page by days, and "URL is on Google" does not prove visibility. The feature inspects only properties the user holds, which rules out checks on competitors; non-owners are pointed to the Rich Results Test or the AMP test.

Quotas are the sharpest dispute. Mike Hardaker, chief executive of Mountain Weekly News, reported on April 25, 2026 that Search Console stopped him after five Request Indexing submissions while he updated 1,192 technical audits, a case PPC Land covered the next day. He argued that the throttle ignores the safety relevance of the content and noted that Bing and DuckDuckGo index his pages without friction. Google's position, as Mueller expressed it, treats a need for frequent manual submissions as a possible sign of quality doubts. Reported limits conflict. Practitioner guides, including one from Alev Digital, put the daily cap near 10 to 12 per property; PPC Land's April coverage cited an industry assumption of several hundred for established sites; Google publishes no figure.

Reliability of the surrounding data is a second problem. The page indexing report sat on a June 11, 2026 data point for more than two weeks, according to Search Engine Land. A comparable delay in November 2025 led Google to say that only reporting was affected, according to Search Engine Journal, which added that the inspector kept working for single addresses. Interpretation is a third problem. When "crawled - currently not indexed" counts rose across many properties in spring 2026, practitioners disagreed over how much was genuine deindexing and how much display error, as PPC Land's Search Console explainer records.

Not the same as

Page indexing report. An aggregate list of indexed and excluded addresses across a property, with reasons. The inspector explains one address; PPC Land's explainer on indexing covers the stages behind both.

Rich Results Test. A public tool that runs on any address without ownership and uses the same Google-InspectionTool fetcher. Google took it out of beta in July 2020.

Indexing API. A push channel limited to job posting and livestream pages, with a default quota of 200 requests, according to Google's quickstart. It is no general substitute for Request Indexing.

IndexNow. An open push protocol launched by Bing and Yandex. PPC Land noted Google's absence from the protocol in 2024, and third-party trackers still reported it as non-participating in April 2026.

Recent developments

Googlebot's per-URL fetch limit fell from 15MB to 2MB on February 6, 2026, a cut of 86.7%. Google's explanations followed: a nine-point crawl overview on March 3, a podcast episode on March 12 describing Googlebot as one client of a shared platform, and two blog posts on March 31 on byte limits and crawler categories. The inspector's help page, as retrieved in October 2026, does not mention the cap.

Ranking volatility spiked on May 13 and 14, 2026, with "crawled - currently not indexed" rising for many sites. In July 2026, on a Search Off the Record episode, Mueller said that when Google's systems have serious quality reservations they reduce crawling and indexing, which can appear as mass "not indexed" statuses, according to Search Engine Journal.

On October 6, 2026, Google grouped Search Central videos into two playlists; the technical track lists effective use of the inspector among its topics.

Timeline

  • May 2014: Google publishes guidance on rendering pages with Fetch as Google, the legacy predecessor
  • June 25, 2018: URL inspection announced in the beta Search Console, with rollout over the following weeks
  • January 2019: HTTP response, page resources, JavaScript logs and screenshots added
  • March 28, 2019: Fetch as Google removed
  • August 2019: Google testing tools adopt evergreen Googlebot rendering
  • July 2020: Bing Webmaster Tools adds a URL inspection beta
  • October 14, 2020: Request Indexing disabled
  • December 2020: Request Indexing restored after 69 days
  • October 11, 2021: Public testing tools aligned with the inspector
  • January 31, 2022: URL Inspection API launched
  • December 1, 2023: Mobile-Friendly Test and Mobile Usability report retired
  • February 6, 2026: Googlebot fetch limit cut from 15MB to 2MB per URL
  • April 25, 2026: Hardaker reports hitting the Request Indexing quota after five submissions
  • June 11, 2026: Page indexing report data stalls for more than two weeks
  • October 6, 2026: Google publishes a technical SEO playlist that includes the inspector

Summary

Who: Google operates the feature inside Search Console; site owners, publishers, technical SEO practitioners and agencies with access to a verified property use it. Daniel Waisberg and Dori Rosenberg announced the API in 2022.

What: A per-address diagnostic reporting Google's indexed copy of a page, its crawl and canonical data, enhancements found, and the outcome of an optional live fetch, with a daily-limited Request Indexing action and an API capped at 2,000 queries per property per day.

When: Announced on June 25, 2018 as the replacement for Fetch as Google, which was removed on March 28, 2019. The API followed on January 31, 2022.

Where: In Google Search Console, through the inspection bar at the top of any screen or the Inspect link beside addresses in reports, and through the Search Console API. Bing Webmaster Tools offers a separate equivalent.

Why: Site owners cannot see Google's index directly. The feature is the one first-party source that explains, for a single address, whether it was crawled and indexed and what blocked it, which makes quotas, stale data and reporting delays matters of commercial consequence.

Editorial notes

  • Skipped workflow step: the publication list placeholder in the project prompt is still unfilled, so the secondary-publication sweep was not run. Search Engine Land and Search Engine Journal appeared through general-web searches and carry the third-party points.
  • Duplicate check: no existing PPC Land "Explaining" article for this term turned up; the adjacent ones (Search Console, indexing, Googlebot) are linked.
  • Conflict inside PPC Land coverage - Request Indexing year: the April 2026 quota article dates the button's return to December 23, 2021. The original PPC Land post is dated December 24, 2020, and Search Engine Land puts the suspension at October 14, 2020 plus 69 days. This article uses December 2020. The quota article's timeline probably needs correcting.
  • Conflict inside PPC Land coverage - names and dates: the quota article names Mike Hardaker and dates his LinkedIn post April 25, 2026 (timeline entry April 26). "Explaining Search Console" calls him Chris Hardaker and gives April 26. This article uses Mike and April 25 for the post.
  • Conflict - daily quota: the quota article cites an industry consensus of several hundred Request Indexing submissions a day; practitioner guides cite roughly 10 to 12; Hardaker reported five. Google publishes no number. All three are attributed, none is verified.
  • Inferred dates: January 2019 for the screenshot and resource additions is inferred from the Search Engine Land article's position and image timestamp, as the text carries no date. The late-June 2026 end of the page indexing delay is not stated in this article for the same reason. The July 2026 Search Off the Record date rests on a third-party summary; Search Engine Journal's text carries no date.
  • Third-party only: the March 28, 2019 removal of Fetch as Google (Bill Hartzer), the 10 to 12 daily figure (Alev Digital) and IndexNow's April 2026 status come from non-Google sources. No introduction date for Google-InspectionTool is given because none could be sourced from Google.
  • Not verified: whether the 2MB fetch limit applies to Google-InspectionTool. The article says only that the help page does not mention it.
  • Time-sensitive claims to recheck before publication: API quotas (2,000 a day, 600 a minute per property; documentation last updated August 28, 2025), the list of 26 rich result types on the help page, and the absence of any published Request Indexing quota.
  • Link check: all PPC Land links were returned by site searches; each was matched against the claim it carries.