If Google Search Console shows a growing pile of URLs marked “Crawled – currently not indexed,” you are looking at pages Google has visited, evaluated and then chosen to leave out of its index. That status is different from a penalty and different from a technical error, and treating it like either one usually wastes time. In most cases, the real explanation is simpler and harder to fix: the content does not offer anything Google’s index doesn’t already have several times over.

What “crawled – currently not indexed” actually tells you

Under the Indexing > Pages report in Search Console, Google separates two very different states. “Discovered – currently not indexed” means Google knows a URL exists but has not gotten around to crawling it. “Crawled – currently not indexed” is further along: Google downloaded the page, ran it through its evaluation systems, and decided not to add it to the database it serves search results from.

Google staff have been direct about this distinction at recent events and in official channels. When Google crawls a page, that’s effectively a download. Whether it gets stored in the index depends on whether Google’s systems judge it useful enough to be worth serving. Two things tend to drive that judgment: whether the page shows genuine personal experience, and whether it contains knowledge that isn’t already sitting on ten other domains.

It’s worth separating this from routine noise in the same report. Feed URLs, paginated archive pages, and parameterized duplicates showing up as crawled-not-indexed is normal and not something to chase — Google is correctly declining to index a non-canonical copy of a page you already have indexed elsewhere.

Two causes, and only one of them is common

A technical block (rare, but worth ruling out first)

Before assuming a quality problem, confirm Google can actually see your content. Open the URL Inspection tool, find the affected page, and run “Test Live URL.” If the rendered result shows a heading and boilerplate but no body content, something is preventing Google from reading the page — not judging it.

A real-world example: a site that had recently migrated platforms had almost every page stuck in crawled-currently-not-indexed. The cause was a robots.txt rule — Disallow: /*?* — meant to block low-value parameter URLs like ?replytocom or ?utm_source. The new theme, however, loaded its CSS and JavaScript through parameterized URLs, so the same rule blocked Google from seeing most of the page’s actual content. Removing the block let pages start reappearing in the index. If your live test shows Google reading full content correctly, a technical cause is unlikely and you should move straight to evaluating quality, which is what’s covered next.

A quality judgment (the common case)

When a page renders fine but still doesn’t get indexed, Google’s own explanation is that it looked at the page and “found it not to be good” relative to what’s already available. That doesn’t necessarily mean the page is badly written. It often means thousands of other pages already say the same thing, and Google’s systems see no reason to add another copy. Google has also acknowledged experimenting with temporarily indexing a page to see whether users respond well to it before deciding whether to keep it in — which explains why some URLs flicker in and out of the index over time.

Commodity content is the pattern behind most of these cases

Commodity content is anything a large number of writers could produce with roughly the same result: a definition, a how-to, a listicle that restates what’s already ranking, assembled without direct experience of the subject. Non-commodity content is the opposite — it carries a first-hand account, a proprietary data point, a documented technical case, or an opinion earned by doing the work, not just researching it.

This distinction maps closely to Google’s published guidance on helpful content, which repeatedly asks the same questions: does the page provide original reporting or analysis; does it add insight beyond the obvious; if it draws on other sources, does it add substantial value rather than just restate them; does it offer more than what’s already ranking for the same query. A page can be accurate, well-formatted and free of errors and still fail every one of those tests.

What Google’s Search Off the Record team said about this directly

Google’s own Search Off the Record podcast tackled this exact report in an episode on reading the Indexing Report, and the discussion is unusually candid for an official Google channel. Asked whether crawled-currently-not-indexed is often a quality signal, the team’s answer was blunt: when Google’s systems have serious concerns about a site’s overall quality, they reduce how much of it gets crawled and indexed. Seeing a pattern of pages in this state across a site is described as the system effectively saying, “we know about this, we’ve looked, and we’re not convinced yet.”

The team also pushed back on the idea that this is purely a technical problem to patch. Their framing was that once you notice a broad pattern — many pages not indexed with no technical explanation — the right response is to step back and assess the site’s quality with fresh eyes, which is difficult precisely because it’s your own work. They flagged AI-generated content specifically: pages that read as competently written but generic, where a visitor can tell “anyone could have written this,” tend to accumulate in this bucket. They also made a point that’s easy to overlook: quality isn’t just about the text. Pages buried under ads, interstitials, and filler content — the kind seen in recipe posts padded with an unrelated life story before the actual recipe — get judged on the full experience, not just the words a scraper would extract.

How to audit pages you believe deserve to be indexed

Pull your list from Search Console’s Pages report under Indexing, filter to crawled-currently-not-indexed, and start with URLs you genuinely believe should rank. For each one:

• Search the queries you’d expect the page to rank for and check whether an AI Overview or similar answer box already fully resolves the query on the results page. If a user gets a complete answer without clicking through, that’s a strong signal your page is competing with a summary Google already generates for free.
• Compare your page against whatever is currently ranking for that query using the same criteria Google’s helpful content documentation lays out: original analysis, insight beyond the obvious, added value versus existing sources, and substance compared to what’s already ranking.
• Ask directly: would a knowledgeable reader learn something from this page they couldn’t get from the first three results already ranking?

Large language models are genuinely useful for this step, not as a substitute for judgment but as a way to get a fast, structured second opinion. Feeding a competing top-ranking page and then your own page into the same prompt, and asking where your version falls short against the same criteria, surfaces gaps faster than manual review alone — provided you treat the output as a starting point for editing, not a verdict.

What actually fixes it

If it’s technical, the fix is straightforward: correct the robots.txt, meta robots, or rendering issue, then request reindexing or wait for the next crawl. If it’s quality, there is no shortcut. Improving a page stuck in this status usually means adding something that genuinely wasn’t there before: a first-hand test, a proprietary number, a documented edge case, an opinion backed by direct experience rather than research alone. Google’s own Quality Rater Guidelines use the word “effort” so often it’s become a shorthand for what separates ranking content from everything else — content that visibly required work a competitor’s version skipped.

Be cautious about scale-content strategies that cover a topic and then systematically expand into every related “fan-out” query the topic implies. Publishing at that volume without proportional added value is a pattern search systems are actively working to suppress, and a drop in organic traffic with no manual action in Search Console is a plausible symptom of exactly this. AI tools can speed up research and drafting, but content built primarily by prompting a model to cover a topic other sites already cover well is, by definition, commodity content — the same problem in a faster wrapper.

This connects to a broader shift worth understanding: as Google has explained about the link between site quality and non-indexed pages, indexing decisions increasingly function as a site-wide quality signal rather than a page-by-page one. It’s also worth checking whether related Search Console reports are flagging real problems or noise — see our breakdown of which Search Console “errors” Google says usually aren’t real problems — and, if you’ve recently requested a fix, our guide on when to actually use Search Console’s Validate Fix button.

Frequently asked questions

What’s the difference between “discovered” and “crawled” – currently not indexed?

Discovered means Google knows the URL exists but hasn’t crawled it yet. Crawled means Google downloaded and evaluated the page and chose not to index it. The second status carries more information about how Google judges the content itself.

Is crawled-currently-not-indexed a penalty?

No. It’s not a manual action and won’t appear as one in Search Console. It reflects an algorithmic decision that the page isn’t worth serving relative to what’s already indexed, which can stem from either a technical block or a quality assessment.

Should I worry about /feed/ or parameterized URLs showing up in this report?

Generally no. Non-canonical duplicates, feeds, and pagination showing up here is expected behavior, not a sign of a content problem on the canonical version.

How long does it take to recover a page from this status?

If the cause is technical, recovery can happen within a few crawl cycles once the block is fixed. If the cause is quality, there’s no fixed timeline — Google needs to recrawl and re-evaluate the improved page, and for pages with genuine content overlap issues, a partial rewrite rather than a small edit is usually what’s required. See also how our take on bot-detection screens dropping pages from the index for a related technical-access failure mode, and Google’s explanation of the related “indexed without content” error if your issue looks like rendering rather than quality.

Does AI-generated content automatically get excluded from the index?

No, but content that reads as generic and interchangeable — regardless of whether AI was involved in writing it — is the pattern most associated with this status. The determining factor is whether the page adds real value, not the tool used to produce it.