What each meta tag actually does
The head of a page fills up with meta tags, and it is tempting to read them as a control panel: set the values, tell the search engine what to do. They are not that. Each tag is a separate mechanism with a separate audience, and at least one of them does nothing whatsoever.
The four lines this tool writes
The Meta Tag Generator takes a page title, a description and a comma-separated keyword list, and emits exactly four lines: a <title> element, a description meta, a keywords meta, and a viewport meta fixed at width=device-width, initial-scale=1. It escapes the characters that would otherwise break your markup, and it does the whole thing in your browser. That is the entire tool. Worth knowing what each of the four is for, because they are not four versions of one idea.
The title element
The title is the page's name, and it is the one entry here that is genuinely load-bearing. It is the browser tab, the default bookmark text, the label a screen reader announces, the text that gets carried into a shared link, and the clickable line in a search result.
One caveat that surprises people: the title you write is a strong candidate for that clickable line, not a guarantee of it. Search engines openly rewrite the displayed title when they judge that some other text on the page — a heading, usually — describes the result better for the query somebody typed. You are supplying the best available version, not filling a slot.
The description
The description meta is a pitch, and only a pitch. It is not an input to ranking; the engines say so plainly, and have for a long time. What it is used for is the grey line of text underneath the result, and even there it is a candidate rather than a fixture: when a passage further down your page answers the query more directly, that passage is what gets shown instead.
So write it for the person reading the result, not for a machine. Say what the page is and what the reader gets. Repeating your title back at yourself wastes the space, and so does a sentence that could sit under any page on the site.
The keywords tag
The meta keywords tag has been ignored by the major search engines for many years. Filling it in does not do anything for them. It was abandoned for an obvious reason: it was a self-reported list, invisible to readers, with no cost to inflating it, so it became noise within about a week of anyone noticing it existed.
The tool still emits the line, for two honest reasons. It is valid HTML, and a handful of smaller engines and internal site-search systems do still read one. Neither is a reason to spend time on it. There is a small cost, too: the tag is public, so it publishes a tidy list of the terms you were hoping to be found for to any competitor who opens view-source.
If you want a keyword to matter to a search engine, the place for it is the visible text of the page — used because the page is genuinely about that thing, in the title and the headings and the prose, at whatever frequency reads naturally. There is no target frequency, and hunting for one produces text that a human can feel has been tampered with.
The viewport tag
This one is not about search at all, which is why it looks out of place in an SEO tool. It is a layout instruction to mobile browsers. Without it, a phone browser assumes your page was built for a desktop-width canvas, lays it out at that width and zooms out, and the reader gets a tiny unreadable version of a page that would have fitted perfectly. width=device-width tells it to lay the page out at the real width of the screen, and initial-scale=1 tells it not to zoom.
It matters, but through the front door: a page that is unusable on a phone is unusable on a phone, and no tag rescues that.
What is not in the four lines
Plenty, and deliberately. A robots meta, which is how you ask for a page not to be indexed, is a different tag with real consequences if you get it wrong. A canonical is a link element, not a meta, and it tells engines which URL of several near-identical ones is the real one. Open Graph and card tags control how a link looks when it is shared into a chat or a social feed, and they are separate tools on this site. None of those is generated here, and a generator that quietly emitted a robots tag you did not ask for would be a hazard rather than a convenience.
How to spend your time
Write the title first, put the distinctive words at the front, and make it describe this page rather than the site. Write the description as one or two sentences you would be happy to see quoted next to a competitor's. Leave the keywords field empty or fill it in twenty seconds, because it will not be read. Then go and improve the page itself, which is the only part of this that anybody — person or crawler — actually reads.