An online SEO audit checks a website against the technical factors search engines rank by. Paste a URL, and the tool runs 400 checks, then returns a 0–100 score, a prioritized issue list and a PDF report. A quick scan takes about 20 seconds. We audited our own site, scored 70.5, and explain what that number means.
What a website SEO audit is
A technical SEO audit is a check of your site against the factors that drive indexing and rankings: server responses, redirects, robots.txt, sitemaps, canonicals, meta tags, structured data, speed, mobile rendering and accessibility. The output is not a pass/fail verdict — it is a list of specific problems with priorities and fix instructions.
A separate layer is AI search visibility: bot policies for GPTBot and ClaudeBot, llms.txt files, and your standing in Bing — the index ChatGPT and Copilot pull answers from. Classic checklists skip this layer; our canon audits it alongside the basics.
How to check your site online in 4 steps
- Paste your site address into the audit form — the homepage is enough, the crawler discovers the rest
- A quick scan returns a preliminary score in ~20 seconds; the full audit adds page rendering and Lighthouse
- Read the breakdown: critical P0 issues come first, and every finding lists affected pages and a fix
- Download the PDF report — hand it to a developer or contractor as a ready task list
What gets checked: 400 points in 36 groups
The auditor runs a canon of 400 checks grouped into 36 sections: HTTP responses, redirects, robots.txt, sitemaps, indexing, canonicals, HTML and meta tags, links, duplicates, Open Graph, structured data, images, performance, mobile, accessibility, security, JS SEO, analytics, monitoring and AI visibility. Every check carries a priority, and the priority sets its weight in the score.
| Priority | Checks | Weight | Examples |
|---|---|---|---|
| P0 — critical | 49 | 10 | HTTPS 301 redirect, robots.txt not blocking the site, canonical resolving to 200 |
| P1 — important | 214 | 3 | Title and description length, self-canonical, Open Graph, error monitoring |
| P2 — nice to have | 137 | 1 | Extended structured data, PWA manifest, Common Crawl presence |
The score is the share of weight you earn: every passed check adds 10, 3 or 1 point by priority, every failed one loses it. Checks that do not apply to your site are excluded from the denominator, so a landing page and an online store are both scored fairly, with no penalty for sections they do not have.
We audited our own site and scored 70.5
Running the tool on yourself is the best test there is. A quick scan of app-kit.ru took 19.6 seconds: 10 pages, 235 automated checks, a score of 70.5 out of 100, with 14 failures. Here is what the auditor found on our own site:
- The http version of the homepage answers 200 instead of a 301 to https — a critical P0, a duplicate host in the index
- robots.txt still carries a deprecated Host directive pointing at an old subdomain
- None of the 10 pages sets a self-canonical
- Titles on 3 pages run past 70 characters, so search snippets get truncated
- No monitoring for 404s, 500s or JS errors
- robots.txt has no explicit sections for AI bots: GPTBot, ClaudeBot, PerplexityBot
The user should get the full audit: an overall score, a breakdown of every point, and a PDF report you can hand to a developer. A cut-down demo never answers the question of what to fix first.
Online audit vs manual checklist vs agency
| Criterion | Online audit | Manual checklist | Agency |
|---|---|---|---|
| Time to result | Seconds to minutes | Hours per pass | Days to weeks |
| Coverage | 400 points, every time | Depends on your focus | Depends on the vendor |
| Repeatability | Identical run, any moment | Fatigue by point 100 | Every report is a new quote |
| Cost | Free | Your time | Project budget |
These approaches complement each other. Automation covers the measurable base and covers it the same way on every run; a human is needed where the tool honestly reports “manual check” — content quality, commercial factors, Search Console and Bing Webmaster Tools panels. Our canon has 8 such points out of 400, and the report lists them separately.
Quick scan vs full audit: what the difference is
The quick scan crawls up to 10 pages without rendering and runs 235 automated checks: server responses, redirects, robots.txt, sitemaps, canonicals, meta tags, markup and AI visibility. Its job is a preliminary score in seconds, before you leave the page.
The full audit adds in-browser page rendering and a Lighthouse pass. That unlocks the performance, mobile, accessibility and JS SEO groups — all 400 canon points. The full run is also what produces the PDF report. It takes a few minutes, so the page shows progress by stage: crawl, render, checks, report.
How to read the result and where to start
Work through the report by the cost of each point, not by list order. Failed P0 checks come first: each is worth 10 points and usually signals an indexing risk for the whole site. Then mass P1 failures — one root cause often covers dozens of pages at once, like a self-canonical missing from a template. Leave P2 for planned releases.
- A failed P0 is a day-one fix: HTTPS redirects, reachable robots.txt and sitemap, canonicals without chains
- A P1 failure with dozens of affected pages is almost always one template fix, not a per-page chore
- High-weight “unknown” checks deserve a manual look before you write the points off
- Under 60 — fix the technical base before touching content; over 85 — your next wins are in content and links
Frequently asked questions
How long does a site check take?
A quick scan takes about 20 seconds: the crawler visits up to 10 pages without rendering and runs the base check groups. The full audit with page rendering and Lighthouse takes a few minutes — the page shows progress, and you can leave it open in a tab.
Is it free?
Yes, with no sign-up. The full audit with a PDF report is computationally expensive, so it is rate-limited per IP to a few runs a day. The quick scan has a softer limit.
What do P0, P1 and P2 mean?
They are the canon priorities. P0 is critical for indexing, weight 10: a failure can keep the site out of search. P1 is important, weight 3: it visibly affects rankings and snippets. P2 is nice to have, weight 1: improvements without direct risk.
How is this different from Lighthouse?
Lighthouse measures a single page: performance, accessibility, basic SEO. The audit looks at the site as a whole — it crawls sections, cross-checks robots.txt, sitemaps, canonicals and duplicates across pages, and covers AI visibility. Lighthouse itself runs inside the full audit as one data source for the performance groups.
How often should I run an audit?
After every meaningful release and at least once a month. Technical debt builds up quietly: a new section without canonicals, a forgotten redirect, broken markup. A repeat run takes seconds and shows the delta against your previous score.
Can I audit a competitor’s site?
Yes — the auditor works with any public URL, which makes it handy for sizing up competitors before entering a niche. The crawler respects robots.txt and keeps a gentle request rate; pages behind a login and private addresses are out of reach by design.
How accurate is the score?
Every finding in the report comes with evidence: affected page URLs, server responses, tag contents. You can re-verify any point by hand. Ambiguous cases get an “unknown” status instead of a failure — the score is never dragged down by the tool guessing.
Next: the checklist and the tools
For a walkthrough of the 18 most important canon checks with real failure examples, read our SEO audit checklist. If the audit points at heavy images, we have dedicated guides and tools for compressing and resizing them.
Join the discussion
Comments are coming to the App-Kit blog soon. Got a correction or a question?
Send feedback