Critical
The six that break pages outright.
- Server errors (5xx)
- Broken pages (4xx)
- Broken internal links
- Redirect loops
- Missing title tag
- Forms that submit over HTTP
The score
A health score with the formula printed next to it.
The sample site scores 82 out of 100 across 1,247 crawled pages, with the issues grouped by category so the widest problem is at the top. The score is 100 − Σ weight × √(share of pages affected), weighted 25 for a critical issue, 8 for a warning and 2 for a notice — and the product shows you that sum rather than hiding it.
- Good is 80+, Needs work is 50+, below that is Poor
- Only pages actually fetched count towards the score
- A cancelled crawl produces no score rather than a misleading one
Templates
Catch a bad template before it multiplies.
Programmatic pages fail in groups, not one at a time. MrKeyword folds crawled URLs into templates and reports each one's page count, the share that is thin, and how alike the pages are — so you can fix the template instead of 400 URLs. The Issues tab groups the rest by category, so the widest problem is at the top of the list rather than whichever check happened to run first.
- Near-duplicate pages grouped by URL template
- Thin-content share per template
- The pages MrKeyword judged most alike, listed for you
- Issues grouped by category and by new, persisting or fixed
With your AI assistant · Example
What did the last audit of this page find, and what do I fix first?
Tools usedget_latest_page_auditget_daily_issues
Performance
See the experience problems search engines can measure.
The audit keeps technical issues and user experience signals together, with field data named when it exists and a clear needs-data state when it does not. That makes the next fix easier to prioritize than a generic speed score.
- LCP, INP and CLS shown with their source
- Mobile and desktop context kept visible
- No invented field data when Google has none
History
Know what improved after the fix.
A single audit tells you what is wrong today. Crawl-over-crawl history shows what is new, fixed and still present, so the audit becomes an operating record instead of a one-off checklist.
- Compare the latest crawl with the previous snapshot
- Separate fixed, new and persisting issues
- Keep the score change beside the underlying checks
The checks
An SEO audit tool with 60 checks.
6 critical, 24 warnings and 30 notices — a full technical SEO audit of the site, with a plain-English explanation of the fix on every one.
Warnings
Twenty-four issues that quietly hold rankings back.
- Duplicate titles
- Thin content
- Slow responses
- Invalid structured data (JSON-LD)
- Mixed content and pages served over HTTP
- Orphan pages and soft 404s
- Missing mobile viewport
- Hreflang without a return link
Notices
Thirty-one smaller things worth cleaning up.
- Images missing alt text
- Multiple H1 tags
- Pages buried deep in the site
- Indexable pages missing from the sitemap
- Render-blocking scripts in the head
- Images without width and height
- Form fields without labels
- Missing security headers
On screen
Five tabs, one website audit.
Overview
A health score out of 100 with its band, counts of critical, warning and notice issues, a crawl-over-crawl diff of what is new, fixed and persisting, and the biggest score deductions.
Issues
Grouped by type, filterable by severity and by whether an issue is new, persisting or fixed since the last crawl, each with its fix and the pages affected.
Crawled pages
Fifteen filters — indexable, noindex, canonicalised, 4xx, 5xx, redirects, blocked by robots.txt, orphan pages, in sitemap, slow responses and more — with status, indexability, click depth, inlinks, word count and response time.
Core Web Vitals
LCP, INP and CLS for mobile and desktop from Google's PageSpeed Insights, with the source named: Chrome UX Report field data, or a lab test where Google has no field data yet.
Audit one page
Fetch any single page on demand for a full rule check: a page score, category meters, ranked recommendations with an effort tag, structured data found and suggested JSON-LD drafts. It does not use your crawl allowance.
Scheduled re-crawls
Weekly, every 30 days, or off. Owners and admins set the schedule, and the next run time is shown.
Each Core Web Vitals panel names whether the numbers are Chrome UX Report field data for that exact URL, field data for the whole origin, or a lab run — because those three are not the same evidence.
How the numbers are made
How the crawler behaves on your server.
The crawl runs from MrKeyword's own infrastructure, politely, and identifies itself so you stay in control.
robots.txt always wins
Parsed to RFC 9309, with Crawl-delay honoured up to 10 seconds. MrKeyword never ignores robots.txt to reach a page, and it only crawls sites in your own workspace.
Two requests a second, maximum
Per site, with a 15-second request timeout. HTML is capped at 5 MB, sitemaps at 20 MB, up to 50 sitemap files and 50,000 sitemap URLs.
Same site only
The www. and apex forms are treated as one site, nofollow links are not followed, and tracking parameters are stripped before a URL is queued.
It says who it is
The crawler identifies itself as MrKeywordBot/1.0, so you can see it in your logs and control it from your own robots.txt.
Web Vitals without invention
Good means LCP ≤ 2.5 s, INP ≤ 200 ms and CLS ≤ 0.1. A lab test cannot measure INP, so where Google has no field data the panel reads “Needs field data” instead of printing an estimate.
Scope
What this does not do
Being clear about the edges is part of the product. These are deliberate limits, not gaps we are hoping you will miss.
- It does not crawl sites you have not added to your workspace, and it never ignores robots.txt to reach a page.
- It does not estimate Core Web Vitals. Where Google has no field data, the panel says so.
- It does not guarantee a ranking change; it reports technical problems and how widespread each one is.
- Structured-data drafts are built only from what is visible on the page — they are a starting point to check in Google's Rich Results Test, not markup to paste unread.
Read next
Related pages.
See it with your own data.
Connect a property and this page becomes your numbers. Free to start.
Get started free No credit card required.