A link copied from an ad, a newsletter or a social post carries more parameters than the page needs. Most of them exist for the platform that sent the visitor, and none of them changes what the page shows. Keeping them has a cost: the same page appears under several addresses, a canonical tag points at the tagged version, and analytics reports look better than the site.
The URL cleaner removes the tracking set, sorts the query and returns the names it dropped.
curl "https://api2.globus.studio/v2/url-clean?url=https%3A%2F%2Fglobus.studio%2F%3Futm_source%3Dx%26a%3D1"The default list
Every utm_* parameter goes. That is the widest convention and covers most ad platforms. The endpoint also removes a set of click identifiers: fbclid, gclid, dclid, gbraid, wbraid, yclid and msclkid, plus the Twitter, TikTok and Instagram identifiers.
Marketing automation adds its own: the Mailchimp pair mc_cid and mc_eid, the HubSpot pair _hsenc and _hsmi, mkt_tok from Marketo, and the pk_* campaign family from Matomo. All of them are in the default list.
The full set is on the documentation page. Anything not on it stays, which is the important part: a parameter named page or id can change the response, and removing it would break the link.
Extra names and sorting
The remove parameter takes up to twenty names, comma separated, for a project with its own convention. A partner that appends ref=newsletter can be cleaned without waiting for a general rule.
Sorting is on by default. Two links that differ only in parameter order become the same string, which helps comparisons and redirect rules. Turn it off with sort=0 when the order matters, which is rare outside of signed URLs.
What normalisation does not touch
Scheme and host are lowercased and default ports are dropped. An empty path becomes /. The fragment is kept unless strip_fragment=1 is passed.
The path itself is not rewritten. A trailing slash can change routing on some servers, and a cleaner that silently removes it creates broken links with a clean appearance. If the path needs a rule, that is a redirect, not a cleanup.
Reading the result
The response returns the cleaned URL, a changed flag and the list of removed names. The flag is the check to run before writing a redirect: a link that comes back unchanged needs no rule, and a link that changes only because of sorting tells you the parameters were already harmless.
Encoding is preserved for values, so a signed or encoded parameter that stays in the link keeps its original form.
Where it fits
Clean links belong in canonical tags, in sitemaps, in redirect targets and in documentation. They do not belong in ad tracking, where the parameters are the measurement. The endpoint is a tool for the places where a URL becomes permanent.
A canonical that keeps a campaign parameter tells search engines that the tagged version is the primary page. Over a few months the site collects several variants of the same article, and reports split between them. Cleaning the link before it reaches the template avoids the problem at the source.
The same logic applies to redirects. A redirect rule that matches a cleaned URL misses every tagged link, and a rule that matches the tagged version misses the cleaner one. Cleaning first and redirecting second gives one stable address.
Keep the original link while a campaign is running. The report needs the identifier that a cleaner removes.
No request is made and nothing is cached. The endpoint is pure string work and sits in the cheapest rate tier.
The catalogue page has the parameter table and a live example: URL cleaner API.