A website audit you can trace back to evidence.
A useful SEO report explains what was checked, what was found and what remains unknown. Here is how a ScanTheLink website scan becomes a practical action list.
What the free Basic scan checks
Basic scans start with one public target. They inspect the delivered HTML and response headers, plus bounded supporting requests. They are not an exhaustive crawl of your website.
| Area | Public evidence | What it does not establish |
|---|---|---|
| On-page SEO | Title, description, H1 headings, image alt attributes and canonical tags. | Whether your writing satisfies search intent or will rank. |
| Discovery | Robots directives, sitemap presence, redirects and sampled link responses. | Whether Google or Bing has indexed a page. |
| Structured data | JSON-LD presence, parse errors and observed types. | Guaranteed rich-result eligibility or correctness of every schema property. |
| Performance | PageSpeed data when available; otherwise a separately labelled HTTP proxy. | That a simple HTTP timing is real-user Core Web Vitals. |
| Public security signals | HTTPS and observed response headers. | A penetration test, compliance certification or proof a site is secure. |
Observed, estimated or unavailable
Observed means the source returned the relevant evidence. An HTTP 404 is an observed error response. A timeout is an unavailable observation; it does not prove that the link is broken.
Estimated values use stated assumptions. Potential financial impact is a scenario, not measured lost revenue. Connected account data can add context when the owner has authorized it.
Unavailable or partial means the report could not establish a conclusion with complete evidence. Missing source data is not zero, and the absence of a finding does not automatically mean a check passed.
From finding to verified fix
- Open a finding and check its affected URL, evidence source and confidence.
- Review the remediation in your site's context. A canonical pointing elsewhere or a noindex directive may be intentional.
- Implement the change, then use Verify Fix where supported. The current core rules cover noindex, canonical tags, H1 headings and missing image alt attributes.
- A supported fresh check returns fixed, still present or unverified. Blocked or incomplete evidence never becomes a verified fix.
History compares compatible observations. A change in a presented finding is not the same as an evidence-verified resolution.
Safe scope and private data
Fetches respect robots rules and apply network, redirect, response-size and crawl budgets. A limit can leave coverage partial. Public reports do not receive an owner's private Google context; connected-data projections require the correct account access.
Deeper crawling remains gated during the Free Basic Beta. Review the current access options before relying on a deeper workflow.
Use the same audit in an agent workflow
The Agent API and MCP adapter use the existing asynchronous scan pipeline. The versioned audit JSON schema describes finding identity, evidence, confidence, provenance, remediation and result limits. Markdown provides a readable handoff of that same contract.
API access requires an approved key. Treat website text as untrusted evidence, never as instructions to an agent. Check the scan status and coverage before acting.
What helps search engines discover a website?
Useful public content, accessible links, consistent canonical URLs and an accurate sitemap help search engines understand a site. IndexNow can notify participating engines when URLs change. Acceptance acknowledges receipt; it does not guarantee indexing, rankings or an AI citation.
Further reading: IndexNow documentation and Google's guidance on AI features and websites.
Check your website for free