There is no public Google backlink checker. The closest thing Google offers is the Links report inside Search Console, and if you search for a "google backlinks tool" expecting a list of every link pointing at your site, that report will not give it to you. Its own documentation says so in the first paragraph.
That is not a reason to ignore it. It is the only backlink data that comes from Google rather than from a third-party crawl, which makes it the reference point for one specific question: did Google find this link. For every other backlink question, you need a different source.
This is a decision guide for that split, built on Google's documented limits, our own export data, and the recent reminder that the report is not always fresh.
Recommendation at a glance
The question you are asking | Source to use | Why |
|---|---|---|
Did Google discover this new link to my page? | Search Console Links report | First-party confirmation, free, tied to your verified property |
What is my full link profile, including links Google has not surfaced? | A third-party index | The GSC report is a documented sample, not a census |
How many links does a competitor have? | A third-party index | Search Console only reports links to properties you can verify |
Is this link nofollow? | The linking page itself | The GSC report does not specify nofollow |
Is the anchor text profile manipulated? | Search Console link text report, then manual review | The report groups anchors by exact match and is sampled |
Should I disavow? | GSC export, then manual verification | Disavow decisions need live-page confirmation before anything is submitted |
If you only read one row, read the second one. Everything else in this article exists to explain why those two sources disagree and why that is expected.
Who can work with Google's own data alone
Three situations where the Search Console Links report is enough, and paying for an index adds cost without adding an answer.
- Small sites with a short link history. If your profile is under a few hundred links, a sample covers most of it. Our own profile fits in this bracket and the report stays readable.
- Confirmation work on your own property. When you build a link and want to know whether Google processed it, only Google can tell you. A third-party crawler knowing about the link proves nothing about Google.
- Official records for a reconsideration request. If you are documenting link removal for a manual action, you want the record Google itself produced, not a vendor export.
Who needs a third-party index instead
Four situations where Google's report cannot answer the question at all.
- Competitive gap analysis. Search Console shows links to properties you have verified. It will never show you a competitor's profile.
- Coverage disputes. When a link you know exists does not appear in the report, you cannot tell whether Google missed it, deduplicated it, or simply did not include it in the sample.
- Large link profiles. Tables cap at 1,000 rows. For a site with a substantial profile, the export is a truncated view.
- Historical and freshness questions. The report includes links Google found over time, including links that may no longer exist, and it does not timestamp its data.
The six limits Google documents itself
This section exists because most confusion about Google's backlink tool comes from expecting a census and receiving a sample. All six of these come from the Links report documentation.
1. It is a sample. "This report isn't a comprehensive list of every link on your site. It shows a sample of internal and external links to help you understand your site's overall link profile."
2. Links are omitted for specific reasons. "Some URLs might be omitted for various reasons, such as non-indexed pages or deduped URLs." Non-indexed linking pages are a common explanation for a link you can see with your own eyes and cannot find in Search Console.
3. Tables cap at 1,000 rows. "Tables are limited to 1,000 rows, so tables may be truncated in larger or more linked-to sites."
4. The list is historical, not current. "The report includes links that Google has found over time. These links may have since been removed, or a page may no longer exist." A GSC link is evidence of discovery, not evidence the link is live.
5. Nofollow is not reported. "The report doesn't specify if a link is marked as nofollow." You have to open the source page to know.
6. Data is grouped by root domain. Protocol, subdomains, and subdirectories are stripped and grouped together. A link from a subdomain appears under the parent domain, which is why your own subdomains can show up as separate "sites" in the table.
One limit sits outside the documentation. The report does not publish a data freshness timestamp, and in August 2026 it stopped updating for roughly a month without an official explanation. We wrote up that gap separately, including what Google did and did not say: Search Console Links report data frozen since August 8.
Each documented limit removes a question from the report's range. The useful move is matching questions to sources rather than demanding one source answer everything.
What our own export actually looked like
We pulled our own backlink profile on September 11, 2026 to make the decision concrete rather than theoretical. A third-party index reported:
- 1,131 backlinks
- 23 referring domains
- 6 referring domains with only nofollow links
- 0 broken backlinks
- A spam score of 4
- First link seen June 18, 2026
The headline number, 1,131 links, is the one most people would report. It is also the least informative number in that list, and here is why.
Referring domain | Links | Referring pages | Links per page | What it is |
|---|---|---|---|---|
Domain A | 615 | 615 | 1.0 | Template: one link on every page of the site |
Domain B | 346 | 346 | 1.0 | Template: one link on every page |
Domain C | 102 | 102 | 1.0 | Template: one link on every page |
Domain D | 10 | 3 | 3.3 | Partly editorial, some pages linking more than once |
Domain E | 4 | 1 | 4.0 | A single directory listing page |
Domain F | 3 | 1 | 3.0 | One editorial mention |
Read the last column, not the second. Three of our top six referring domains have a links-per-page ratio of exactly 1.0, which is the signature of a site-wide template link: one link on every page the crawler visited. One domain contributed 615 links that way. Editorial placements in the same profile sit at three or four links from a single page.
This is the analysis the ratio gives you for free, and it is the reason a raw link count is a weak metric. Your profile is not 1,131 links. It is a small number of editorial placements plus a larger number of template links, and the two do completely different things.
Same profile, two readings. The ratio separates template links from editorial links, which the link count cannot.
Costs, risks, and blind spots
Search Console. Free. Official. Sampled, untimestamped, silent on nofollow, capped at 1,000 rows, and silent on competitors. Its unique value is that it is the only source that reflects Google's own discovery.
Third-party indexes. Paid, and useful for exactly the questions GSC cannot answer. Every index has its own crawl gaps, so two vendors will report two different referring-domain counts for the same site. That is normal, and it means vendor numbers should be read as estimates with a consistent methodology rather than as truth.
Spam scores. Heuristic, and they vary by vendor. Treat a spike as a prompt to look at the linking page, not as a verdict. Our own score of 4 means nothing on its own; what matters is that the domain with a spam score of 13 contributed 346 template links while the editorial placements sat at 0.
The shared blind spot. Neither source tells you whether a link sends traffic, whether the linking page is indexed, or whether the link is still on the page today. That work has to happen.
The verification loop that makes an export useful
An export from Google is a starting list, not a conclusion. Five checks per row, and the first four are mechanical.
- Is the linking page still live? Request it and check the status code. A 404 explains a link mismatch immediately.
- Is the link still on the page? Fetch the HTML and search for your URL. Historic listings outlive removed links.
- Is it followed? Read the anchor's rel attribute. Google will not tell you, so check the source.
- Is the linking page indexed? An unindexed linking page is one of the documented reasons your link is missing from the report. We wrote the check for that separately: how to check if a backlink site is still indexed.
- Is it template or editorial? Links divided by referring pages. Near 1.0 with a large denominator is a template link.
Rows that fail checks one to three go on an ignore list. Rows that pass and look editorial go into a placement log with the date and the anchor used. Rows that pass but carry a concerning pattern go on a watch list with a review date, not straight to the disavow file. Disavow remains a narrow tool with real caveats, and we covered the decision path in how to disavow toxic backlinks.
If you want the deeper diagnosis workflow once your sources are settled, our Codex backlink diagnosis guide walks through the profile-level analysis. The loop itself is mechanical, which makes it a reasonable agent task. One export, five checks per row, one classified output file. What the agent should not do is decide whether a link is good, because that judgement needs business context. Our SEO agent guide covers where that line sits.
Auspia view: stop asking which backlink tool is accurate. Two sources that disagree are two instruments, not a contradiction. Use Google to confirm discovery, a paid index to understand coverage and competitors, and a verification pass before you act on either.
The next step
If you have never compared the two, do it once on a small scale. Export the Links report from Search Console, pull the same property from a third-party index, and put the referring-domain lists side by side. Differences in coverage, nofollow handling, and duplicate counting will be obvious within twenty minutes, and you will stop treating either number as complete.
If your profile is small, do the ratio analysis on your top ten referring domains first. It usually changes which link you consider your best, and it costs nothing.
The survey evidence supports paying attention here rather than chasing volume. In the Zyppy expert survey published September 9, 2026, backlinks placed second out of more than a hundred candidate factors, nominated by 54.8% of 131 practitioners. Quality and relevance drove that ranking, not count.
FAQ
Does Google have a backlink checker? Not a public one. The Links report in Search Console is the closest equivalent and it is limited to properties you have verified. It shows a sample of links rather than a complete list.
Why does the Search Console Links report show fewer links than Ahrefs or Semrush? Because it is a sample by design, and because it only includes links Google has processed. Third-party tools run their own crawls and report what they found. Both numbers are estimates of different things.
Why can't I see nofollow links in Search Console? The documentation states the report does not specify whether a link is nofollow. You have to inspect the linking page's HTML to determine it.
Why does a link disappear from the report? Three common reasons: the link was removed from the page, the linking page is no longer indexed, or the URL was deduplicated. Google's documentation names non-indexed and deduped URLs explicitly.
Is the Links report still updating? It went roughly a month without a refresh starting around August 8, 2026. Before relying on it for time-sensitive work, check whether the newest links you have built appear. If they do not, treat the data as historical until it refreshes.
How many links do I need before I should pay for a tool? There is no threshold. A better trigger is the question you are asking. The moment you need a competitor's profile, a coverage estimate, or nofollow status at scale, the free report has run out of range.
Author: Iris Campbell, Editorial Evidence Analyst with 2,500+ Sources Reviewed at Auspia. Iris writes about evidence quality, source comparison, and how to tell a measurement from a claim.
