URL parser

The query string is everything after ?. It is sent to the server and is normally used for filtering, searching or tagging where a visit came from.

The fragment is everything after #, and the browser keeps it to itself - it is never sent to the server. That is why it is used for jumping to a section of a page, and why some sign-in services hand a token back through it: the value never reaches the server logs.

They are the five parameters utm_source, utm_medium, utm_campaign, utm_term and utm_content, added to a link so analytics tools can tell where a visitor came from. They exist purely for measurement and never change what the page shows.

Strip them when you just want to send someone a clean link, because leaving them on means that person's visit gets counted against an old campaign. This tool labels UTM parameters so they are easy to spot - hit the bin icon on each row and copy the URL underneath.

A URL can only carry a narrow set of characters, so spaces, accented letters and special characters have to be percent-encoded: each byte becomes a % followed by two hex digits. %20 is a space, and chào becomes ch%C3%A0o because "à" takes two bytes in UTF-8.

The parameter table shows the decoded text so it can be read, with the encoded form on the small line underneath each cell. One special case: inside a query string a + also means a space by web-form convention, so hello+there reads as "hello there".

Because it is normalised the same way a browser normalises it: the host is lowercased, the default ports :443 and :80 are dropped, an empty path becomes /, and spaces and + are rewritten as %20.

The meaning does not change - the server still receives exactly the same values. If you need every character preserved as you typed it, use your original string; the "Rebuilt URL" box is there to give you the result after you edit parameters.

It is the nested parameter notation that PHP, Laravel, Rails and many other libraries understand: empty square brackets mean "append an item to an array", and brackets with a word inside are a sub-key. The "Parameters as JSON" box at the bottom of the page rebuilds that structure so you can see what the server actually receives. A name repeated several times, such as a=1&a=2, is collected into an array too.

The whole analysis runs in JavaScript inside your own browser. The page never sends the URL to any server and never stores it; close the tab and it is gone.

The thing to watch is the "Copy share link" button: that link carries your full URL, so whoever you send it to can read all of it. The tool flags parameters with names like token, password or api_key so you remember to strip them first.

It appends your URL after the # of this page, giving something like nguyenlap.net/en/tools/parse-url#https://.... Anyone opening that link lands straight on the analysis without pasting anything. And because nothing after # is sent to a server, the URL you are looking at never reaches this site's logs either.

Yes. With no protocol, the tool adds https:// for you and says so. The ambiguous cases are handled properly too: localhost:8080/api is read as a hostname with a port rather than a protocol called "localhost", and //cdn.example.com/a.js is read as protocol-relative.

The standard sets no limit, but in practice it is worth staying under roughly 2,000 characters. Some older browsers and a fair number of servers cut off around that mark, and a very long URL pasted into a message or an email is easily broken in two. The character count sits in the top right of the "Structure" box; dropping a few tracking parameters usually shortens things considerably.