Compare dated scans
Use /diff for changes to scores, checks, pages, readable content, raw HTML and captured screenshots. The homepage overlay is free; other pages are paid.
Open before / afterThe whole journey
From a hidden issue
to a reviewed change.
Follow the work in order. We explain what each tool does, what you can act on and where the evidence stops.
Live
Run a free audit
THE PAGE / TWO READINGS
Illustration
What you see
A page that
A page that
looks finished.
What a crawler may be told
<meta name="robots"
content="noindex">
What that instruction means
Do not include this page in the index.
01
Live
Crawlers read your site differently than you do.
A page can look complete while a robots rule blocks a crawler, a noindex instruction prevents indexing, or a canonical points elsewhere. Search bots and training bots also serve different purposes.
THE PAGE / TWO READINGS
Illustration
What you see
A page that
A page that
looks finished.
What a crawler may be told
<meta name="robots"
content="noindex">
What that instruction means
Do not include this page in the index.
What this does not tell youCrawler access rules do not establish visits or citations. Read our original crawler policy study.
02
Live
Scan your site. See what was inspected.
Start with one public URL. We inspect up to 150 pages and distinguish inspected pages from discovered URLs. Free views show category scores and selected findings; other details are sealed. Anonymous visitors get 1 fresh scan per site per month; a free account gets 2.
01
Fetch
Read public pages and supporting files.
02
Evaluate
Apply checks where inputs are available.
03
Report
Show evidence, coverage and limits.
What this does not tell youUninspected pages are not passes. Unavailable inputs do not become successful checks.
03
Live
Know what the score does and does not say.
Findings are grouped Gate, Core, Amplifier and Hygiene, worst first. The current score is an equal-weight fraction of measured conditions. AEO, GEO, Local and privacy observations receive no numeric grade.
187
registered detectors
The full catalog
25
scoring-eligible evaluators
14 distinct conditions
Varies
conditions measured per scan
Applicability + captured inputs
Of the scoring-eligible evaluators, 18 cover SEO, 4 accessibility and 3 security. The title condition is counted once. The other 162 detectors produce unscored observations.
Our own site · Saved report #670 · 17 September 2026 · Legacy method evidence-v2004
Discovered is not inspected.
1
page inspected
249
URLs discovered
#670
17 Sep 2026
Uninspected pages are not passes. This saved report cannot certify the other 248 URLs.
A historical policy score.
79Saved policy score, legacy method evidence-v2004, not a rescan, not a ranking.
The original report recorded 69 passes, 5 warnings and 65 informational rows. Today’s measured-conditions method is different.
Two warnings are visible.
The excerpts below use the saved report’s own words. The historical sample must stay scoped to pages actually inspected; it does not establish the structure of the full site.
corpus.site_structure · saved wording, excerpt This saved scan inspected 1 page and discovered 249 URLs.
core.trust_pages · saved wording, excerpt 1 of the 3 trust basics are missing.
Sealed means not displayed.
72 rows are sealed: 3 warnings and 69 passes. Their totals are visible, but the findings and evidence are not disclosed in this free view.
◇ 1 sealed warn rows · Core
◇ 1 sealed warn rows · Amplifier
◇ 1 sealed warn rows · Hygiene
◇ 8 sealed pass rows · Gate
◇ 12 sealed pass rows · Core
◇ 11 sealed pass rows · Amplifier
◇ 38 sealed pass rows · Hygiene
The report’s own “not measured” list.
- Whether you actually rank
- Authority & backlinks
- Content depth & topical coverage
- The competition above you
- Indexation reality
- Real Core Web Vitals & brand corroboration
These are historical dimension labels, not a claim that those connections are available in this sample.
Saved snapshot included for offline reading.
Open saved report #670What this does not tell youThe score does not measure rankings or AI citations. This legacy sample is not a v2006 rescan.
04
Live
One repair. Then the next.
The guide brings forward a repair with plain explanation, generic steps, copy buttons and a developer handoff. “I did this” is a session-local acknowledgement, not verified or durable history.
Illustration with example.com data, not a customer result.
01 / Found
Missing canonical
The example page has no canonical tag. First decide which URL should represent this content. A canonical is a preference signal, not a command to Google.
02 / Choose platform
Generic guidance
Choose how you edit the page. These are generic steps, not WordPress, Shopify, Wix, Squarespace or Webflow tutorials.
03 / Make the change
One step you can review
1. Confirm that https://example.com/ is the preferred URL.
2. Add this tag inside the page’s <head>.
3. Save and publish the change.
<link rel="canonical" href="https://example.com/">
04 / Acknowledge
“I did this” is your note
An acknowledgement records that you followed a step in this session. It is not evidence that the condition passes, and it is not durable repair history.
05 / Recheck
Read the page again
The actual guide fetches fresh evidence for one supported condition on one page. This illustration makes no scan request. Choose an outcome to see both paths.
This condition passes on the page we rechecked
Only the named condition and page are confirmed. Other affected pages still need their own evidence.
Not confirmed: the page still fails or could not be inspected
Review the current evidence. A failed fetch or an unknown result never becomes a pass.
What this does not tell youPlatform choice still leads to generic steps. Platform-specific tutorials and automatic site edits are not built.
05
Live
Check the change, not the hope.
Recheck one of 11 supported page conditions with fresh evidence: title, H1, meta description, canonical, canonical conflicts, noindex, HTML language, viewport, mixed content, HTTPS or multiple titles. The limit is 3 attempts per 5 minutes per account and site. Rechecks do not update the saved report.
60-day receipts
A baseline, weekly snapshots and a receipt from stored scans and real Search Console rows. The workflow has connection and delivery limits.
Read the data requirementsWhat this does not tell youFailure or unknown evidence never confirms a repair. Use a full rescan for other conditions. Google’s response is not integrated.
06
Needs connection
Readiness is not results. Here is where results come from.
When connected, read-only Search Console supplies actual search observations. Authority is a graph snapshot; Core Web Vitals need available data; paid competitor context depends on a third-party feed.
Search Console
When connected, Google’s read-only data supplies clicks, impressions and average positions. A valid property connection is required. Illustration: Not measured
Read scope and limitsAuthority snapshot
Domain-level authority and linking-domain count from our Common Crawl graph index. This is a dated snapshot, not individual backlinks. Illustration: Not measured
Read scope and limitsCore Web Vitals
CrUX field observations and PageSpeed lab results appear only when the feed returns data for the domain. Illustration: Not measured
Read scope and limitsCompetitor context
The paid panel depends on a third-party SERP feed and monthly caps. Missing feed data is not evidence of no competition. Illustration: Not measured
Read scope and limitsThis is a connection map, not your account status. No search metrics have been loaded into these tiles.
What this does not tell youNot measured does not mean zero. Missing connection data does not support a performance claim.
07
Partial
Keep watching one site.
Monitor costs $29/mo for one site. The Monday 13:00 UTC scheduler is designed to scan weekly, with emails on a score drop of 3+ points or a new gate failure. Review history and fresh exports when you return.
MONITOR / WEEKLY CADENCE
Illustration
-
Scheduled site scan
Capture a new report for the monitored site.
-
Compare with the prior report
Look for a score drop of 3+ points or a new gate failure.
-
Return to the evidence
Review changes, report history and export files.
Example dates show the intended cadence. They are not completed scans or delivery receipts.
What this does not tell youThe intended monitoring cadence is not a reliability promise or a measure of search outcomes.
08
Live
Hand off one repair, or the whole report.
Copy, download or print one repair brief. Paid exports provide the PDF, action plan, JSON fix file and manifest. The report prompt uses findings visible to your account; share links expose the free view. Agency covers five sites for $99/mo.
Read the report
A saved report in a document you can review and share.
Follow the action plan
Prioritized findings with context, affected pages and suggested next steps.
Keep the evidence attached
Fields include report_id, scanned_at, issues, evidence and suggested_fix.
Format illustrations based on the export contract, not screenshots of a generated export. The pack also includes a manifest.
What this does not tell youGenerated files and model drafts need review before publishing. These tools never write to your website.
09
Live
Choose the scope you need.
Start free, buy one report or keep monitoring the site. All plans use the same scoring method.
Free
$0 one public site scanCategory scores and selected findings.
Explore FreeFull Report
$10 once · one siteDetailed findings, affected pages and export files.
Explore Full ReportMonitor
$29 /mo · one siteFull reporting and scheduled weekly monitoring.
Explore MonitorAgency
$99 /mo · five sitesSeparate client reports and monitoring per site.
Explore AgencyMonitor and Agency stop at limits, no overage. You can cancel a monthly plan from your dashboard. Full Report has no monthly renewal.
What this does not tell youA paid plan opens report detail and the listed workflow. It does not establish better rankings, traffic or AI citations.
10
Live
Useful before and after buying.
Use a focused tool to inspect a condition, prepare a change or compare evidence. The Learn hub links the fix library, glossary, engine guides and research.
Inspect
Check canonicals, structured data, crawler access and more.
Browse the 14 free toolsLearn
39 fix guides and 31 glossary entries, counted from the published files.
Open the guide libraryRead the research
Robots.txt policy across 653 reachable domains from the Tranco top 1,000, for 15 named crawlers. Measured 5 July 2026.
Read the original studyWhat this does not tell youAccess rules do not establish visits or citations. A generated snippet is a proposal you must review.
Before you start
A few honest answers.
Can a score change when the method changes?
Yes. Saved report #670 uses legacy evidence-v2004. New reports use measured-condition scoring from v2006. Compare scoring versions and dated evidence before treating score movement as a site change. Read the details.
Does AuditLamp predict rankings or AI citations?
No. The scan does not measure rankings or AI citations. A pass is evidence about a condition, not an outcome. Connected Search Console can show observed performance when available. Read the details.
Does AuditLamp fix my website for me?
No. You or your developer make the change. Generic repair steps, generated files and AI-written drafts all require review before publishing. Read the details.
How does a recheck decide whether a repair passes?
It fetches one page again and evaluates one supported condition. Fresh passing evidence can confirm that scope. Failure or unknown evidence never confirms a repair. It does not update the saved report. Read the details.
Choose by the task
You may need a different tool.
For keyword research, backlink prospecting or rank tracking, choose a tool built for that job. Evaluate our evidence and handoff on a site you know before buying.
Read the comparisonBegin with the evidence