Free Meta Tag Generator: Build Accurate HTML Head Tags
A meta tag generator is a form with fields.
A meta tag generator is a form with fields. You paste in a page title, a description, the canonical URL and sometimes a social image, and it returns an HTML snippet for your `<head>`. That is the whole product. It does not write your copy, it does not make a page rank, and it cannot tell whether the URL you pasted is live.
We still use them on almost every site we ship, because the value is in the checklist rather than the markup. A decent generator nudges you to fill fields people forget: `rel="canonical"`, the viewport tag, `og:image:width` and `og:image:height`. It also keeps attribute quoting and self-closing syntax consistent, which matters when three people edit the same template.
What it cannot do is decide intent.
What it cannot do is decide intent. Only you know whether a page should be indexed, whether it duplicates another page, or whether it should point at a different canonical. Treat the generator as a typo-proofing layer, not a strategy tool.
Start with the essentials: `<title>`, `<meta name="description">`, `<meta name="viewport" content="width=device-width, initial-scale=1">`, a canonical link, and `<meta charset="utf-8">` at the top of the head. Add `<meta name="robots" content="noindex">` only when you genuinely want a page out of search results — and check twice before you ship it.
The keyword meta tag is still output by plenty of generators.
The keyword meta tag is still output by plenty of generators. Google has said for years that it ignores it, and Bing treats it as a spam signal at worst. It is harmless clutter, but we strip it.
Everything else is situational. `hreflang` for multilingual sites, `theme-color` for mobile browser chrome, and author or publisher tags if your CMS already handles them. If a tag does not change how a page is displayed, indexed or shared, ask what it is doing in your head at all.
The reference pages worth bookmarking are the [MDN documentation for the meta el
The reference pages worth bookmarking are the MDN documentation for the meta element and Google's own guidance on how snippets are generated.
Text length in metadata is measured in pixels, not characters, because the font and viewport matter. Google truncates according to available width, so there is no hard character limit. In practice we aim for roughly 50–60 characters in a title and 150–160 in a description, then measure.
A practical habit: write the description first, run it through a [word and chara
A practical habit: write the description first, run it through a word and character counter, and trim until the key phrase and the reason to click both survive inside the first 155 characters. On a recent internal tool page, our draft description ran to 214 characters and buried the useful part after the ellipsis point. Cutting to 152 kept the promise and the brand name visible.
Remember that Google rewrites descriptions often, pulling text from the page body when it thinks something better fits the query. A tight description raises your chances of being used; it never guarantees it.
Open Graph tags drive previews on Facebook, LinkedIn, Slack, Discord and most me
Open Graph tags drive previews on Facebook, LinkedIn, Slack, Discord and most messaging apps. The must-haves are `og:title`, `og:description`, `og:url`, `og:type` and `og:image`, plus `twitter:card` set to `summary_large_image` for X. The specification lives at ogp.me.
The two mistakes we see most: relative image URLs, and oversized images. `og:image` must be an absolute URL, including the domain and protocol. And a 3 MB PNG will simply fail to render on some platforms or load too slowly to be fetched. Our rule of thumb is 1200×630 pixels, under a few hundred kilobytes, JPG or PNG — and if you are working from a design export, a tool like this photo compressor to 40 kB gets you there in seconds.
Platforms cache previews aggressively.
Platforms cache previews aggressively. After changing tags, use the Facebook Sharing Debugger or LinkedIn Post Inspector to force a refetch, otherwise you will spend an afternoon debugging a change that already worked.
Suppose you are publishing a landing page at `https://example.com/pricing`. Open a generator, paste that exact URL, and fill in: title "Pricing — Example" (short, branded), description of about 150 characters explaining what is on the page, type `website`, and your 1200×630 image URL.
Copy the output into the head of your template, above the closing `</head>` but
Copy the output into the head of your template, above the closing `</head>` but after charset and viewport. Then verify: view source in the browser (or inspect the rendered DOM, since some frameworks inject tags via JavaScript that crawlers may not execute), and run the URL through a sharing debugger. If a tag is missing from the debugger output, it is missing for real.
One edge case that catches people out: single-page apps that set metadata client-side. Search engines generally render JavaScript, but many social scrapers do not, so Open Graph tags set only in the browser will not appear in previews. Server-render them or inject them at build time.
Reused titles and descriptions across a whole site are the classic problem — cat
Reused titles and descriptions across a whole site are the classic problem — category pages, paginated pages and tag archives all inheriting one default. That is a content-management job, not a generator job, but the generator output is where you notice it. The same applies to leftover placeholder text like "Your Company Name Here", which happens more often than anyone admits.
Accidental `noindex`, canonicals pointing at a staging domain after migration, and duplicated tags (one from the CMS, one hardcoded in a template) round out the list. When tags are duplicated, most crawlers take the first occurrence, which is rarely the one you edited.
Our maintenance habit is simple: metadata lives in structured CMS fields rather
Our maintenance habit is simple: metadata lives in structured CMS fields rather than free text, and we spot-check a sample of pages each quarter against a preview debugger. Nothing here guarantees rankings or rich results, and platform behaviour changes without notice — so if search visibility matters commercially, confirm current guidance in Google's and Bing's own documentation rather than trusting any tool's output blindly.
If you just need somewhere to draft, count and check metadata without installing anything, FreeOnline.fyi runs free browser-based tools with no sign-up and no upload of your data to a server you do not control. There are more practical guides on the blog if you want the surrounding workflow, from image compression to copy length.