🔎 Website Audit Report
Paste a page's HTML and get a branded audit PDF: 30 checks on how the page is found, shared, trusted and turned into enquiries, each one written as what it costs rather than as a code. Free, no signup, and your pasted HTML stays in this tab.
Step one: copy the page's HTML
A browser is not allowed to fetch someone else's page for you, so nothing here goes and gets it. You copy the markup once, and everything after that happens on this page.
- Open the website you want to review, then press F12 (or right click the page and choose Inspect).
- Go to the Elements tab (Firefox calls it Inspector) and scroll to the very top line, the one that starts <html.
- Right click that line, choose Copy, then Copy outer HTML.
- Paste it into the box below.
Why the Elements panel and not View Source. View Source shows the markup before any JavaScript has run. On a site built with React, Vue, Webflow interactions or a page builder, that source is close to empty, so an audit of it would report no headings, no images and no contact details on a site that has all three. Copy outer HTML gives the page as the visitor actually sees it. If you cannot open the developer tools, saving the page with Ctrl+S and dropping the .html file below works too.
Step two: the page
Include https:// or http://. The address heads the report, resolves any relative links found in the page, and is the only way to check the connection and mixed content.
Nothing you paste is uploaded, the audit runs in this tab. To be exact about the rest: the page loads its PDF writer from a public CDN and counts an anonymous visit, and the optional note drafter at the bottom sends the findings list, never your file.
Step three: your branding (optional)
PNG or JPG is safest. An SVG is converted to an image here first, which does not always work. Saving keeps your name, contact line and colour in this browser only, never the logo and never the audit. Forgetting deletes that stored copy and leaves what is typed above untouched.
Fix these first
Every check
Take it away
Two pages for most sites: the score, the breakdown and the priority fixes on the first, every check on the second. A site with a lot wrong runs a page or two longer, because the findings are not trimmed to fit. The JSON carries the same results, with an importance on every check and a severity on the ones that did not pass.
Optional: draft the note that goes with it
Pressing this sends the website address and the list of findings to the free drafting service so it can write a short covering note. Your pasted HTML is never sent. If the service is busy, a plain version is written here instead.
A website audit you can hand to somebody
Most free website checkers give you a screen of scores and a signup wall. This one gives you a document. Put in the address, paste the page markup, add your name, colour and logo, and you get a branded PDF that opens with a score out of 100, breaks that score into five areas, then lists what to fix first and what each problem is actually costing in plain sentences a business owner will understand. There is a findings.json alongside it if you want to keep the results or push them somewhere else.
The same report works three ways. Run it on your own site and you have a to do list ranked by how much each item matters. Run it on a client's site and you have the covering document for a proposal. Run it on a stranger's site and you have the one thing a cold approach needs, which is proof that you looked. Report writing services charge from about thirty dollars a month for the white label version of this, and audit gigs sell for a hundred dollars and up.
What the thirty checks look at
Found on Google. Title tag and its length, meta description and its length, how many H1 headings there are, whether a long page has any subheadings, the canonical link and where it points, whether a noindex instruction is quietly keeping the page out of search altogether, and the language attribute.
Shared links. The Open Graph title, description and image that decide what a pasted link looks like in Facebook, LinkedIn, Slack and WhatsApp, plus the Twitter card tag.
Business details. Structured data, whether it actually declares an Organization or a LocalBusiness, whether any of the blocks are broken JSON that Google will discard, a visible phone number, a tap to call link and a clickable email address.
Getting enquiries. Whether there is an enquiry form at all, whether a live chat widget is present, and whether an analytics tag appears anywhere in the markup.
Speed, mobile and safety. The mobile viewport tag, whether the address is https, assets still loaded over plain http on a secure page, images with no alt text, images with no width and height, scripts in the head that block the first paint, an end of life jQuery, how many third party scripts the page pulls in, and the weight of the markup as it stands in the browser.
What it does not do, stated up front
This reads one page of markup. It does not load the site, so it cannot time a real page load, run Lighthouse, follow links, crawl other pages, test whether a form actually delivers, check backlinks, or look at anything a script adds after the copy was taken. The analytics check reads the markup only, so a tag injected by a tag manager or by a plugin will not show up, and the report says so rather than accusing the site of having no analytics. There is no favicon check, because most sites serve one by convention with no tag in the markup and the absence of a tag would prove nothing. The jQuery check stays silent unless the script address itself carries a version number. Because you are pasting the page as the browser has built it, the markup counted here is bigger than the file the server sent, so the size figure is a rough signal and never a measure of download time. Every one of those limits is printed in the PDF too, so nobody reading the report is misled about how it was made.
Getting an accurate report
Use Copy outer HTML from the Elements panel rather than View Source. View Source is the markup as it left the server, before JavaScript has built anything, and on a modern site that can be an almost empty page. If what you paste looks too thin to be a real page, this tool refuses to score it rather than producing a report full of false accusations, which is the single worst thing a document like this can do to your credibility. Include the full address with https:// so the connection and mixed content checks can run, and audit the page you actually want to talk about, usually the home page or the main service page.