Skip to main content
ponderloop

URL Parser

A URL breaks into a scheme, optional credentials, a host, an optional port, a path, a query and a fragment, and the URL Standard (url.spec.whatwg.org) says exactly how. Pasted into the parser, the example reads as protocol https, host bücher.example (sent as xn--bcher-kva.example), port 8443, 2 path segments, 6 query parameters and the fragment section-2. It also carries a username and password, which is flagged as a risk, and 3 tracking parameters (utm_source, utm_medium, fbclid); Clean URL removes them and leaves https://demo:pa%40ss@xn--bcher-kva.example:8443/docs/caf%C3%A9%20menu?q=hello+world&tag=a&tag=b#section-2. Edit any part or row and the URL is rebuilt live. Nothing is fetched or sent: the URL is only parsed as text, in your browser.

Updated Runs in your browser: nothing you enter is uploadedHow this tool works

Stays in your browser

The URL is only read as text on this device. It is never opened, fetched or uploaded, and it is not saved: reload the page and it is gone.

Only used when the URL above is relative (for example ../page?x=1): it is resolved against this address.

Split into 2 path segments and 6 query parameters.

Parts

Edit any field and the rebuilt URL below follows. Fields hold the percent-encoded text, as the URL standard writes it.

A special scheme in the URL standard. Its default port is 443.

Decoded: demo. Credentials in a URL are a risk.

Decoded: pa@ss. Credentials in a URL are a risk.

ASCII form shown. Unicode: bücher.example

Port 8443. The default for https is 443.

2 segments. Decoded: /docs/café menu

Never sent to the server.

Query parameters

One row per name=value pair, in order, repeated keys kept. Values are shown decoded; edit them as plain text.

  • 1
  • 2
  • 3
  • 4
    tracking
  • 5
    tracking
  • 6
    tracking

Tracking parameters found: utm_source, utm_medium, fbclid.

Rebuilt URL

The parser’s normalized form: it can lowercase the scheme and host, convert an international host to punycode, drop a default port and percent-encode unsafe characters.

Risks and encoding problems

  • RiskThis URL carries a password before the @ sign. Credentials in a URL are copied along with the link into browser history, bookmarks, server and proxy logs and chat messages, and RFC 3986 section 3.2.1 deprecates the user:password form for that reason. Send secrets in a header or a form instead, and treat any that appeared here as exposed.
  • NoteThe host bücher.example has non-ASCII letters, so it is sent as xn--bcher-kva.example (punycode). Letters from other scripts can look like Latin ones; compare the punycode form when you need to be sure of the site.

Path segments

#DecodedAs written
1docsdocs
2café menucaf%C3%A9%20menu

Origin: https://xn--bcher-kva.example:8443

The example URL, part by part

What the browser’s URL class returns for the example. Hostname is the ASCII (punycode) serialization.
PartValueNote
ProtocolhttpsA special scheme in the URL standard. Its default port is 443.
UsernamedemoDecoded: demo. Credentials in a URL are a risk.
Passwordpa%40ssDecoded: pa@ss. Credentials in a URL are a risk.
Hostnamexn--bcher-kva.exampleASCII form shown. Unicode: bücher.example
Port8443Port 8443. The default for https is 443.
Pathname/docs/caf%C3%A9%20menu2 segments. Decoded: /docs/café menu
Hash (fragment)section-2Never sent to the server.

Tracking parameters removed by Clean URL

The tool’s own list. A name ending in _ matches every parameter that starts with it. Matching ignores case.
ParameterMatchesAdded by
utm_Any name starting with utm_Google Analytics campaign tags (utm_source, utm_medium, utm_campaign, utm_term, utm_content)
mtm_Any name starting with mtm_Matomo campaign tags
pk_Any name starting with pk_Matomo and Piwik campaign tags (pk_campaign, pk_kwd)
hsa_Any name starting with hsa_HubSpot Ads
fbclidThis name onlyFacebook and Meta click ID
fb_action_idsThis name onlyFacebook
fb_action_typesThis name onlyFacebook
fb_sourceThis name onlyFacebook
gclidThis name onlyGoogle Ads click ID
gclsrcThis name onlyGoogle Ads
dclidThis name onlyGoogle Display and Video 360 click ID
gbraidThis name onlyGoogle Ads (iOS app campaigns)
wbraidThis name onlyGoogle Ads (web to app campaigns)
_gaThis name onlyGoogle Analytics cross-domain linker
_glThis name onlyGoogle Analytics cross-domain linker
msclkidThis name onlyMicrosoft Advertising click ID
twclidThis name onlyX (Twitter) Ads click ID
ttclidThis name onlyTikTok Ads click ID
li_fat_idThis name onlyLinkedIn Ads
epikThis name onlyPinterest Ads click ID
irclickidThis name onlyImpact (affiliate) click ID
igshidThis name onlyInstagram share ID
igshThis name onlyInstagram share ID
yclidThis name onlyYandex Direct click ID
mc_eidThis name onlyMailchimp e-mail recipient ID
mc_cidThis name onlyMailchimp campaign ID
_hsencThis name onlyHubSpot e-mail
_hsmiThis name onlyHubSpot e-mail
hsctatrackingThis name onlyHubSpot call-to-action tracking
mkt_tokThis name onlyAdobe Marketo e-mail token
vero_idThis name onlyVero e-mail
vero_convThis name onlyVero e-mail
oly_enc_idThis name onlyOmeda e-mail
oly_anon_idThis name onlyOmeda e-mail
wickedidThis name onlyWicked Reports
ef_idThis name onlyAdobe Advertising
s_kwcidThis name onlyAdobe Advertising
_openstatThis name onlyOpenStat (Russian ad networks)

Resolving relative references

Each reference resolved against https://example.com/a/b/c.html?x=1#top by the URL parser. An absolute URL ignores the base.
ReferenceResolves to
d.htmlhttps://example.com/a/b/d.html
./d/https://example.com/a/b/d/
../dhttps://example.com/a/d
../../dhttps://example.com/d
/dhttps://example.com/d
?y=2https://example.com/a/b/c.html?y=2
#fraghttps://example.com/a/b/c.html?x=1#frag
//other.example/phttps://other.example/p
https://other.example/xhttps://other.example/x

Default ports

The parser drops a port that equals its scheme’s default, so https://example.com:443/ is written https://example.com/.
SchemeDefault port
filenone (no host port)
ftp21
http80
https443
ws80
wss443

How to parse a URL

  1. Paste the URL. Paste a full address into the URL box. The parts update as you type; nothing is requested from the address.
  2. Read the parts. Check protocol, credentials, hostname (ASCII and Unicode), port, path and fragment, and read the risks and encoding problems listed under the box.
  3. Edit the query table. Change a key or value, add a row, remove a row or move it up or down. Repeated keys stay as separate rows. The rebuilt URL changes with every edit.
  4. Clean the URL. Press Clean URL to remove common tracking parameters such as utm_source and fbclid, then copy the rebuilt URL.
  5. Resolve a relative link. Enter a base URL, then paste a relative reference such as ../page?x=1 to see the absolute URL it resolves to.

Frequently asked questions

What are the parts of a URL?
Scheme (https), then // and optional credentials (user:password@), host, optional :port, path, optional ?query and optional #fragment. The URL Standard defines the parser that splits them, and browsers follow it, so this tool uses the browser’s own URL class.
Why is a username or password in a URL a risk?
The credentials travel wherever the link goes: browser history, bookmarks, server and proxy logs, chat messages and screenshots. RFC 3986 deprecates the user:password form, and a link such as https://example.com@other.test/ really goes to other.test, which phishing relies on. Use a header, a token or a login form instead.
What is punycode and why does the host change?
Host names are sent as ASCII. A name with other letters, such as bücher.example, is converted to xn--bcher-kva.example. The tool shows both forms, because the ASCII one is what the network uses and the Unicode one is what people read. Letters that look alike across scripts can imitate another site, so compare the ASCII form when it matters.
What is the default port?
Each special scheme has one: 80 for http and ws, 443 for https and wss, 21 for ftp. When a URL names that port the parser drops it, so https://example.com:443/ becomes https://example.com/. A different port, such as 8443, is kept.
Are repeated query keys and the order of parameters kept?
Yes. The table has one row per name=value pair in the order written, so tag=a&tag=b gives two rows. Rows you do not touch keep their original text; a row you edit is written in form-urlencoded style, where a space is +.
Which parameters does Clean URL remove?
Common tracking parameters: everything starting with utm_, plus click and campaign IDs such as fbclid, gclid, msclkid, mc_eid and igshid. The full list is in the table on this page. Parameters a site may need, such as ref, id or page, are never removed. Check the result for sites you depend on.
How does a relative URL get resolved?
The parser combines the reference with the base URL: ../d goes one directory up, /d starts at the root, ?y=2 keeps the base path and replaces the query, and #frag changes only the fragment. An absolute URL ignores the base. The table above shows each case.
What percent-encoding problems are reported?
A % that is not followed by two hex digits (%ZZ), bytes that do not form valid UTF-8, text encoded twice (%2520), lowercase hex digits (%c3), and characters you typed raw, such as a space or é, that the parser had to encode. Each message names the part it is in.
Does the tool open or fetch the URL?
No. The URL is parsed as text, locally. There is no request to the address, no redirect check and no upload, so it is safe to paste a link you do not trust. The page records only that a button was used, never the URL.

How it works

Parsing, relative resolution and normalization use the browser’s URL class, which implements the WHATWG URL Standard (https://url.spec.whatwg.org/), so the result matches what a browser does with the same text. The ASCII host is what URL.hostname returns; the Unicode display is decoded by a small Punycode decoder written for this page (RFC 3492). The query table is built from the pairs in the URL, one row each, keeping duplicates and order; an untouched row is written back exactly as it was, and an edited row is serialized the way URLSearchParams does. Rebuilding joins the fields and rows into a string and parses it again, so the shown URL is always the parser’s own serialization. Percent-escapes are checked with the same strict UTF-8 decoder as the URL Encode & Decode tool. The tracking list is written for this project, not taken from a library; it names widely documented ad, e-mail and analytics parameters only. The risk notes are typical cautions, not a security scan: this tool never contacts the address, so it cannot say whether a site is safe.

Sources

More: Web, URLs, images and networkAll developer toolsAll toolsHome & DIY calculators