Find your website's weak spots. Know what to fix next.

All of them. The scan analyses your publicly available site (HTML, HTTP headers, robots.txt, sitemap, structured data), so it doesn't matter whether it runs on WordPress, Shopify, Webflow, Wix, Magento or a custom build. No plugin, no access needed.

The free instant scan usually finishes in under 30 seconds and can be run 10 times per day per domain. The full audit with every module and the AI action plan takes 30–120 seconds.

No. We analyse the publicly available website and its measurable technical, content and structural signals. We do not judge what you sell, how you price it or how you run your company, and we do not change your business model.

No. The recommendations explain how your website can present your existing business more clearly to people and machines. That might mean a clearer page title, better structure, verifiable company information or structured markup. This cannot guarantee a recommendation, but it can remove website-level barriers.

We do not ask for a business plan, internal documents, passwords, financial information or a customer database. The system reads what is already publicly available on the website you submit and produces website-level findings from it.

No. The scan is read-only: it fetches and analyses public pages and technical responses. The report suggests improvements, but it never signs in, writes data or applies fixes.

Among other things, we inspect HTML, HTTP headers, robots.txt, sitemaps, headings, metadata, structured data, public copy, performance data and the presence of buyer-decision elements. The report separately identifies anything we could not measure reliably.

No. The public URL is enough. We do not ask for CMS or hosting login details, admin passwords, database access or any other confidential internal data.

Every finding from all 6 modules (SEO, GEO/AI readiness, security, performance with Core Web Vitals, conversion coverage/UX, product pages), a concrete fix instruction plus the raw evidence for each one (HTML snippet, HTTP header, robots.txt line, JSON-LD type or measured value), a prioritised action plan (quick wins, medium, long term), the "Questions your customers ask" chapter, a data-coverage indicator, an AI executive summary, a shareable report link and a downloadable PDF.

A website's health keeps changing: a theme update, a new plugin, an oversized image or an expiring certificate all cause regressions. We re-scan automatically every month, chart your score over time and email you when it drops. Automatic monthly re-scans never consume your manually started audit allowance.

SEO helps you rank in Google's results. GEO (Generative Engine Optimization) helps ChatGPT, Perplexity, Google AI Overviews and Gemini cite your site. That needs different things: allowed AI crawlers, llms.txt, clean factual copy and strong structured data. We check both.

More and more buyers no longer scan Google results — they ask an AI and choose from the 2–3 sources it recommends. If your robots.txt blocks AI crawlers or your pages carry no parsable structure, you're invisible in that traffic.

Yes. Payments are handled by a built-in, PCI-DSS compliant payment provider; card details never touch our systems. You receive an invoice for every purchase.

No. All we need is the URL. The scan looks at your site from the outside, exactly like Google or a visitor does, and changes nothing.

The SEO module checks what Google can actually see and index: title and meta description, heading structure, canonical tags, robots.txt, sitemap, redirects, broken links, image alt texts, internal linking and mobile friendliness. If these are wrong, your pages either never rank or rank for the wrong queries – you end up paying for traffic you could get for free.

The GEO module checks whether AI assistants can read and cite your site: are the search/discovery crawlers (OAI-SearchBot for ChatGPT Search, PerplexityBot, Claude-SearchBot, Googlebot) allowed in robots.txt, is there an llms.txt, is your content factual and well structured, are your Organization/Product/FAQ JSON-LD blocks valid. Buyers increasingly ask an AI instead of browsing search results – if the AI can't read you, you are invisible in that channel. We also check the entity markup field by field (name, logo, sameAs, contact data, @graph linking), the on-page trust signals (About, imprint, company identifiers, author and dates), local NAP readiness on local or service sites, and the paid report adds a score-neutral AI-based web reputation estimate.

HTTPS and certificate state, HTTP security headers (HSTS, CSP, X-Frame-Options, Referrer-Policy), mixed content, exposed admin or backup paths, leaked version information and cookie flags. A weak configuration invites session hijacking, form injection and fake checkout overlays – on a store that means lost orders and lost trust.

We measure real Google Core Web Vitals data on mobile and desktop: LCP (how fast the main content appears), INP (how fast the page reacts to a tap) and CLS (layout shifting). Every extra second of load time measurably reduces conversion, and these metrics are also a Google ranking factor.

Whether the page makes buying easy: a clear headline and value proposition, a visible call-to-action, pricing information, trust signals (contact details, terms, reviews, return policy), form friction and mobile layout. We deliberately call this "conversion coverage" rather than a conversion score: we measure whether the elements a buyer needs are present — we do not test your live cart or checkout flow, and we never predict a conversion rate. You get concrete, ranked suggestions – not vague design opinions.

In the PRO report we analyse up to 10 product pages, 25 on PRO Monitor and 100 on Agency: Product JSON-LD (price, availability, reviews), title and description length and uniqueness, product image count, size and alt texts, canonical tags and internal links. This is what decides whether your products show up in Google Shopping-style results and in AI answers.

Start with Free to see where you stand. If you want to fix things, PRO gives you the full report with step-by-step instructions – that is enough for most sites. PRO Monitor is for stores and sites that change often, where a regression must be caught early. Agency is for agencies auditing many client sites every month.

The PDF report can be downloaded in 5 languages: English, Hungarian, Spanish, French and German. The report framework (section titles, category names, priority labels, the explanations of why each area matters, footer and legal notice) is fully translated into all five. The detailed findings and the AI action plan are written in the language of the scan (Hungarian or English), so on a Spanish, French or German report those parts appear in English.

PageSpeed and Lighthouse are excellent at measuring raw speed, but they do not check AI-search readiness (llms.txt, OAI-SearchBot), structured product data (JSON-LD), security headers or conversion elements. We bring these 5 areas together in one place, and we don't just give you numbers: you get a concrete, prioritised action plan so you know exactly what to fix and where.

They are not interchangeable, and that is the most common misunderstanding. OAI-SearchBot drives ChatGPT Search results — if it is blocked, your site cannot appear in ChatGPT answers. PerplexityBot powers Perplexity answers and citations. Claude-SearchBot drives visibility inside Claude's answers. Googlebot powers Google Search and AI Overviews. The training group — GPTBot, ClaudeBot, Google-Extended — is used for model training and other AI uses, so blocking it is a legitimate business decision: it does not affect your ChatGPT Search visibility, and Google states that Google-Extended does not affect inclusion or ranking in Google Search. ChatGPT-User is the fetch that happens when a user pastes your link into ChatGPT; if it is blocked, that link will not open. The report shows, for every one of these bots, the exact line in your own robots.txt that decided the status, plus a copy-paste ready fix.

Free: 10 instant scans per day per domain (and 20 per IP address). PRO: the one-off purchase gives 1 credit for a single full audit on any domain. PRO Monitor: 3 monitored domains, 30 manually started full audits per month, and the 3 slots can be swapped 2 times in total. Agency: up to 25 monitored domains, 100 manually started full audits per month, 5 domain swaps per month. Automatic monthly re-scans never consume the allowance on any plan, while scans started in the app and via the REST API share the same allowance. The allowance resets at the start of every month.

The Agency plan API key lets you start a full audit from your own system, CRM or automation, list your audits, fetch a single audit summary and the full report JSON (findings, scores, AI action plan). You can manage monitored sites (list, add, remove) and query your white-label branding. PDF download, a dedicated trend endpoint and webhook notifications are not part of the API yet.

Every failing finding ships with its measurement evidence: the exact HTML snippet, the HTTP response header, the decisive robots.txt line, the JSON-LD type we found, or the measured value (word count, image format, LCP). You don't have to take our word for it — your developer can verify the same thing on the same page in a minute. The evidence appears both in the web report and in the PDF.

Starting from your page copy and site type, the system identifies 8–10 real buyer questions (delivery time, warranty, sizes, who is behind the company) and then checks whether your copy answers them. There are three states: answered (with a quote from your page), partial, and unanswered. "Answered" only stands when the answer can be quoted verbatim, so it is not guesswork. For the unanswered ones we suggest which page or section should cover it — that pays off in both search and AI answers.

The report states which modules measured real data and where no measurement was available (no Chrome UX Report data for the domain, no product page found, or only limited measurement because of bot protection). In those cases you get a "not assessable" state instead of an invented score. For the same reason the conversion part is coverage, not a score: we measure whether the elements a buyer needs are present, and we do not test your live cart or checkout flow.

No. Our engine behaves like an external visitor and uses AI-bot User-Agents (e.g. GPTBot) to read public data and server headers. It does not write to your database, fill out forms, start checkout flows, or overload your server. The request will appear in your server log like any other visit.

The diagnosis is strictly based on the official guidelines of Google, Schema.org and modern answer engines. The report is built from live server responses, so it contains objective facts: if JSON-LD, canonical tags, security headers or AI-crawler access are missing, your visibility on modern platforms is measurably reduced. You are not giving your developer an opinion — you are giving them a prioritized task list with code-level fix instructions that can be checked against the official specifications.

Yes. We don't stop at "is there JSON-LD" – we check the entity markup field by field: does the Organization / LocalBusiness node carry name, url, logo, description, sameAs and contact data, and do the blocks form a connected graph (WebSite node, @id references, BreadcrumbList). AI engines use exactly these fields to work out which real company they are talking about; with missing fields the markup exists but links the entity to nothing. Every finding ships with the JSON-LD snippet we found as evidence.

E-E-A-T (experience, expertise, authoritativeness, trust) is about whether a real, identifiable party stands behind the content. We measure the signals that are visible on the site: sameAs / external profile links, an About page, imprint and company identifiers (VAT, registration number), contact and privacy pages, author plus publish/update dates on content sites, and outbound references to sources. On top of that, the paid report includes an AI-based web reputation estimate with named sources: it is explicitly an estimate, it does not affect your score, and when there isn't enough data it is simply omitted.

If the site is local or service-style, a dedicated module checks local readiness: whether the visible name, address and phone match the structured data (NAP consistency), the key LocalBusiness fields, opening hours, a map / directions link, and how easy the contact details are to find. Note that we do not query the Google Business Profile itself – we measure what is on your site, because identical data on profile and website is what lets Google and AI engines connect them. On a non-local site this module is marked "not applicable" and never reduces your score.

Every paid report includes 3 Verify Fix runs. A run re-executes the machine checks — served HTML, rendered DOM, robots.txt, AI bot access, schema and the sampled product pages — and marks each finding as fixed, still failing, or a new issue, with fresh measurement evidence next to every claim. It is not a full audit: PageSpeed and the AI analysis do not run, so it produces no new overall score, and it consumes no credit or audit quota. If the site blocks us during a run, we flag it as a limited measurement and report nothing as fixed.

Yes. Each paid report includes 3 competitor comparisons, for up to 3 competitor domains you choose. We run the same machine checks on their site, then show check by check and by category what they pass and you do not — and the other way around. It is not a score contest: we produce no combined score, and anything that was not measurable on both sides (Core Web Vitals, or a bot-protected site, for example) is never listed as their advantage.

We suggest domains based on organic search overlap — sites that rank for keywords similar to yours. The suggestion is only a starting point: you can dismiss any of them as not relevant, and you can always add a competitor manually. If automatic suggestions are unavailable at that moment, manual entry still works.

On PRO Monitor and Agency, yes: for monitored sites we measure weekly across 5 assistants (ChatGPT, Gemini, Perplexity, Copilot, Claude) whether your site appears in the answer to 10 real buying questions, and whether it gets recommended. You get a 0–100 visibility score, a question-by-engine matrix, a trend, and which competitor is recommended instead of you. Up to 5 of the questions can be your own, and you can track up to 3 competitors. Assistant answers are sampled and shift over time, so this is a measurement, not a guaranteed ranking, and it does not affect your audit score.

Use the WhatsApp button in the bottom-right corner to send us a message. Ask away on WhatsApp – we reply as soon as we can.