SEO Tools

Eight tools, in two kinds. Four check something you have already written — a title, a description, a body of copy — and tell you what a search engine or a reader will do with it. Four produce an artefact you can paste somewhere: a block of meta tags, a robots.txt, a tagged URL, an indexing request.

None of them can tell you where you rank. Ranking needs a live crawl of the whole web, and nothing running in a browser tab can do that. What these tools remove is the mechanical class of mistake — the ones entirely within your control and entirely preventable.

Titles are measured in pixels, not characters

Meta Title Checker exists because “keep it under 60 characters” is approximate advice that fails in a predictable direction. Google truncates by rendered width, not by count. The desktop budget is roughly 580 pixels at about 20px Arial and mobile roughly 680, and since a capital W is several times as wide as a lowercase i, two titles of identical character length can differ by a hundred pixels.

The tool measures your actual string at those widths and shows where the cut falls. It also marks the 30-character point, because that is around where a phone result starts clipping — and most searches happen on a phone.

The practical consequences: put the distinguishing words first, think hard about brand suffixes on pages that need their whole title for the topic, and never assume a title is shown in full. Whatever you write should still read sensibly when it is cut short.

Descriptions have a different budget

Meta Description Checker works the same way with different numbers — about 920 pixels on desktop at 14px Arial and 680 on mobile at 12px Roboto, with a character ceiling near 158. It reports how many lines your text fills before clipping, which is the figure that actually matters, because a description cut mid-word looks careless in the results.

Worth holding onto: the description is not a ranking factor. It is an advertisement for the click, and Google rewrites it freely whenever it believes a passage from the page answers the query better. Write it for a person scanning ten blue links, not for a crawler.

Generating tags instead of typing them

Meta Tag Generator turns a title, description, canonical URL, image and audience settings into the complete block — charset, viewport, theme colour, Open Graph, Twitter Card and the canonical link — ready to paste into a head element. Its value is completeness rather than cleverness. The tags people forget are the ones that only matter when a link is shared, and they find out after the post has already gone round.

Keyword density, and why the number misleads

Keyword Density Checker counts how often each word and phrase appears relative to the total word count and lists the most frequent terms with their percentages.

The arithmetic is simple — occurrences divided by total words — and that simplicity is the problem. One appearance of “image compression” on a 1,000-word page is 0.1% even though the phrase contains two words, so phrase densities always look small. Comparing your figure against a round threshold such as “2%” is meaningless. There is no density that ranks a page, and no major search engine has treated one as a signal in well over a decade.

The tool is genuinely useful for the opposite of optimisation: catching the term you have written eleven times in three paragraphs, which reads badly to a person and looks like manipulation to everything else. Use it as an editorial mirror.

Campaign URLs are a discipline, not a form

UTM URL Builder assembles tagged links from the five standard parameters — utm_source, utm_medium, utm_campaign, plus the optional utm_term and utm_content — and an internal utm_id.

The parameters are the easy part; the vocabulary is the discipline. If one newsletter is tagged utm_source=newsletter in March and utm_source=Newsletter_March in June, your analytics reports two channels and neither number is trustworthy. Two habits fix most of this: keep a short controlled list of source and medium values and never coin a new one for a single campaign, and stay lowercase with hyphens, because the values are case-sensitive in most platforms while looking as though they are not.

Crawl control

Robots.txt Generator builds the file from user-agent rules, Disallow and Allow paths, an optional crawl delay and a sitemap reference. Two details are worth internalising.

Inside a disallowed area the most specific rule wins, so Disallow: /private/ followed by Allow: /private/public-page does exactly what it appears to. And crawl-delay is honoured by Bing and Yandex while Google ignores it completely, using Search Console's own setting instead — a delay of ten seconds slows some crawlers and does nothing at all to the one you probably care about.

robots.txt governs crawling, not indexing. A page blocked here can still be indexed if other sites link to it, and will then appear with no description whatsoever. Keeping a URL out of the index needs a noindex directive on a page that is not blocked from crawling.

Getting a new page seen

Instant Indexing is the only tool in this category that talks to the outside world, so it is worth being precise about what it can do.

The Google tab calls Google's Indexing API, the only push mechanism Google exposes. Google documents that API for pages carrying JobPosting or BroadcastEvent structured data; it accepts other URLs and answers 200 for them, but promises nothing about whether they get crawled. The cap is 200 publish requests per project per day and there is no bulk endpoint — one URL per request. The calling service account must be an Owner of the property in Search Console, and its credential is pasted into your browser and used there.

The IndexNow tab is the wider-reaching half: one protocol accepted by Bing, Yandex, Naver and Yep, with no daily cap and up to 10,000 URLs per request.

Neither can press Search Console's Request Indexing button on your behalf. That button has no API behind it.

Detecting machine-written prose

AI Content Detector measures structural signals — how much sentence length varies, how often paragraphs open with a connective, the density of a list of phrases machine writing overuses, em-dash frequency — and reports each against its own ceiling. Those terms are independent evidence and deliberately do not add up to the headline score.

It is a heuristic and cannot know who wrote anything. A careful human who writes uniformly long sentences scores as machine-like, while padded copy with a varied rhythm scores as human. Treat it as a description of your prose's surface patterns, which is what it measures, rather than a verdict.

What this category cannot do

There is no rank tracker, no backlink index and no site audit here, because all three need crawling and storing the web, which is server-side work. These eight tools are the pre-publish checklist: does the title fit, does the description read well when cut, are the tags complete, is the campaign URL consistent, is the crawler allowed in, and has the page been announced. Every one of those is a mistake you can still fix.