Is it actually free?
Yes, and there is no email gate. You get the full report on screen, including every individual check and the evidence behind it. We would rather show the work than describe it.
HomeFree audit
( Free SEO and GEO audit )
Audit any live page in about fifteen seconds.
We fetch your page, your robots.txt and your sitemap, measure twenty things across crawlability, structured data, AI crawler access, content and technical health, then rank what to fix. Every check shows the evidence behind it. No account, no email, nothing stored.
- Checks run
- 20
- AI crawlers read
- 6
- Turnaround
- ~15s
- Always
- £0
( What the report covers )
Six layers, and every one of them measured, not guessed.
R.01
Crawl and indexation
Status, redirects, canonical, meta robots, and whether the page can be indexed at all.
- Self referencing canonical check
- noindex detection
- Redirect chain
R.02
AI crawler access
Your robots.txt parsed against six named AI crawlers, because a blocked crawler cannot cite you.
- GPTBot, ClaudeBot, PerplexityBot
- Google-Extended, Applebot-Extended, CCBot
- Wildcard rules resolved
R.03
Structured data
Every JSON-LD block parsed, with the schema types it declares listed back to you.
- Types actually found
- Open Graph presence
- Markup versus visible content
R.04
Content depth
Word count, heading structure, text to markup ratio, and image alt coverage.
- Body word count
- H2 and H3 structure
- Alt text coverage
R.05
Technical health
Measured server response, HTML weight, HTTPS, viewport, and declared language.
- Time to first byte, measured
- HTML payload size
- Mobile viewport
R.06
Discovery
robots.txt and XML sitemap located, fetched and counted, plus internal link volume.
- Sitemap found and counted
- Declared in robots.txt
- Internal link count
( How it works )
Four steps. No email gate.
The scoring runs in code, from what the crawl actually observed. Claude reads those measurements and ranks the fixes. It is never asked to supply a number, because a score a model invented is not a score.
STEP 01
Drop in your URL
No account and no email. We normalise what you type, so yourcompany.com is enough.
Instant
STEP 02
We fetch it live
The page, its robots.txt and its sitemap are fetched from our servers and parsed.
About 5 seconds
STEP 03
Everything is measured
Twenty checks across four pillars, each scored from what was actually observed rather than estimated.
Immediate
STEP 04
Claude ranks the fixes
The measurements go to Claude, which orders them by what will move the result most. It reads the data; it never supplies a number.
About 10 seconds
( What this does not do )
We will not tell you what ChatGPT thinks of you.
Plenty of audit tools now report an AI visibility score across four assistants, a share of voice against your competitors, and a count of how often each model cites you. Ask how it was measured and the answer is usually that it was not.
We do not query those systems, so we do not report on them. What we measure instead is the part you actually control and the part that has to be right before any of it matters.
If GPTBot is disallowed in your robots.txt, no amount of content strategy will get you quoted, and that is a one line fix most sites have never checked. If your page carries no structured data and two headings, there is nothing for a retrieval system to lift.
Every number in your report came from a request we made to your site while you waited. You can check any of them yourself, and we show you the evidence so you can.
( Questions )
What people ask about the audit.
Do you store my site or my results?
No. The page is fetched, measured, scored and returned to your browser. Nothing about your site is written to a database, and there is no account to create.
Can you tell me how ChatGPT or Gemini describe my brand?
No, and be careful of any tool that says it can. We do not query those systems, so we do not report on them. What we do measure is whether their crawlers are allowed to read your page at all, which is the part you control and the part that has to be right first.
Are the scores real, or generated by the model?
Real. Every score comes from a measurement taken during the crawl, calculated in code. The model is given those measurements and asked to rank the fixes. It cannot change a score and it is never asked to guess at one.
Why does it only audit one page?
Because a single page can be measured honestly in fifteen seconds, and most problems show up on the homepage anyway. A full site crawl is a different job with a different cost, and we would rather do the small thing properly than the big thing vaguely.
What if my score is bad?
Then you have a ranked list of what to fix and can work through it yourself. Everything in the report is something your own developer can action. If you would rather we did it, the call is fifteen minutes.
Want the fixes done rather than listed?