SEO Audit Tool

Eight areas: technical SEO (HTTPS, redirects, canonical tags, indexability), titles, meta descriptions and headings, crawlability (robots.txt and the sitemap), links (broken, redirected and thinly linked), content (thin or duplicated pages), images (alt text, dimensions, lazy loading, formats), structured data, and mobile and speed basics. Every issue lists the affected pages, why it matters and how to fix it.

Each check has a weight, more for serious problems. A page-level check counts the share of pages that pass it; a site-level check either passes or does not. Each category is scored from its checks, and the overall score is a weighted average of the categories, capped when there are critical problems such as a site that blocks search engines entirely. It is a measure of how many well-established problems a site has — not a prediction of where you will rank, which no outside tool can know.

Always. The crawler identifies itself as FreePandaBot, reads robots.txt before fetching anything, and stays out of anything it disallows. If your robots.txt blocks the bot entirely the audit stops and says so, and how to allow it. It fetches a few pages at a time with a pause between them, and never logs in, submits a form or runs scripts. There is more on the bot’s page.

To keep it free and fair to the sites it visits, an audit crawls a limited number of pages — the report says how many, and whether the limit was reached. Signing in raises the limit. The pages it samples are the ones nearest your home page and those your sitemap lists, which is where most problems show up first.

No. It reads the HTML your server sends, which is what most search crawlers see first and what most AI crawlers see at all. If a page is an empty shell that a script fills in, the audit says so — that is one of the most consequential findings it can make.

No. They are estimates from HTML and response headers, measured from our server: page weight, compression, render-blocking scripts, the viewport tag, response time. Real Core Web Vitals come from a browser measuring real loading, which this does not do. Treat the speed findings as a checklist of likely causes rather than a measurement.

We store the address you entered and the report, so you can come back to it and share it. Anyone with the report’s link can read it — the link is long and random, and the report contains only what a crawler could see on your public pages. We do not store anything about you beyond a one-way hash of your network address, used to limit how many audits one person can start.

Yes, for public pages — it sees only what any visitor’s browser would. Sites can opt out by blocking FreePandaBot in robots.txt, and we honour that. Repeat requests for the same site within a few minutes are answered with the recent report rather than crawling it again.

The most common reasons: robots.txt blocks the crawler, the address does not resolve, the server refused the connection or took too long, or the site’s security certificate is invalid. The message on the failed audit says which. Only public websites can be audited — addresses on private networks, and any address with a username or password in it, are refused.