URL parameters glossary
Every tracking parameter LinkClean strips by default — plus a few functional ones (like hl) everyone wonders about. One-line summary per entry; tap for the long version.
UTM campaign tags
- utm_sourceutm_source names where a click came from — “newsletter”, “twitter”, “google”. It's a Google Analytics campaign tag, not part of the page. Removing it never breaks the link.
- utm_mediumutm_medium names the marketing channel — “email”, “social”, “cpc” (paid search), “organic”. Pairs with utm_source to answer “which channel?”. Removing it never breaks the link.
- utm_campaignutm_campaign labels the marketing campaign — “spring-launch”, “black-friday-2026”. It buckets clicks inside Google Analytics so the publisher can compare campaigns. The page renders identically without it.
- utm_termutm_term names the paid-search keyword that triggered an ad click — “running shoes”, “privacy app ios”, etc. Only appears on paid-search URLs (Google Ads, Bing Ads). LinkClean strips it everywhere; the destination page never reads it.
- utm_contentutm_content names the creative variant or CTA that was clicked — “hero-button”, “v2-banner”, “footer-link”. Publishers use it to A/B-test which version of an ad or email works. The destination page never reads it.
- utm_idutm_id is GA4's campaign-ID UTM tag — a numeric identifier added by Google Analytics 4 to join clicks to a specific campaign row. Newer than utm_source / utm_medium / utm_campaign; same safe-to-strip property.
Referral & source
- sisi is a share-identifier token both YouTube and Spotify add to links copied from their apps. It ties the click back to the sharing session. LinkClean strips it host-scoped to youtube.com / youtu.be / spotify.com so a parameter called si on an unrelated site is left alone.
- CMPCMP is the campaign code the Guardian adds to its article links.
CMP=share_btn_urlmeans the link was copied with the share button, and other values name the app or the newsletter it came from. The article loads the same without it, and LinkClean strips it on theguardian.com. - sX (formerly Twitter) adds two share-tracking parameters to outbound links: t= (the per-share session token) and s= (which surface the share came from — iOS app = 20, web client = 19, third-party = 46, etc). Both are host-scoped to x.com / twitter.com in LinkClean and stripped by default.
- guccounterguccounter is one of three parameters left on Yahoo and Engadget links after Yahoo's consent page sends you on, with
guce_referrerandguce_referrer_sig. Yahoo has not explained them anywhere this page could find, butguce_referrerdecodes to a web address, which reads as the page you came from. The article loads the same without them, and LinkClean strips all three.
Ad-click identifiers
- gclidgclid is Google Ads' per-click identifier — added to every outbound click from Google Search ads, YouTube ads, and the Display Network. It exists to credit ad spend. Removing it never breaks the link.
- gbraidgbraid is the post-ATT replacement for gclid in iOS app-install flows. Google Ads added it in 2021 after Apple's App Tracking Transparency made the classic cookie-based gclid unusable on iOS. It attributes ad clicks → app installs without needing a cross-app identifier.
- gad_sourcegad_source is one of Google Ads' newer auto-tagged URL parameters (rolled out late 2023 alongside GA4's enhanced conversions). Encodes a numeric source identifier (
gad_source=1= Search Ads, etc.). Strip it like the rest of the Google Ads tail. - wbraidwbraid is the post-ATT replacement for gclid on iOS web destinations. Same era and motivation as gbraid (rolled out 2021), but used when the destination is a website rather than an app. Aggregates attribution at a campaign level instead of per-user.
- dcliddclid is Google's display-and-video click identifier — the DoubleClick / Display & Video 360 (DV360) analog of gclid. Where gclid is for Search Ads, dclid is for display-network and YouTube video ads.
- srsltidsrsltid is Google Shopping's search-result listing ID — added to every outbound click from the Shopping tab (paid and free Product Listings alike). Identifies which result panel was clicked, useful only for Google's conversion bookkeeping.
- gad_campaignidGoogle Ads' numeric campaign ID, added to ad clicks beside gad_source. Open the gad_campaignid page for what the number is and why it's safe to strip.
Yandex
- yclidyclid is a per-click ad identifier — same role as gclid (Google) or msclkid (Microsoft). Two independent ad platforms attach the same name: Yandex Direct (dominant on Russian-language search) and Yahoo! JAPAN Ads, which auto-tags ad click-throughs with it. Strip it before sharing.
- ysclidysclid is a click identifier that Yandex Search appends to the address of a result you click. It lets the destination site's analytics tie the visit back to a Yandex search click. The page loads the same without it, and LinkClean strips it on every site.
Analytics IDs
- sca_esvsca_esv is an undocumented token Google Search adds to its own results URLs, first widely noticed in 2023. Google has never said what it means. The search runs identically without it, and LinkClean strips it on Google's search domains.
- vedved is Google Search's encoded description of a link on the results page — which link it was, what kind, and where it sat. It was decoded by outside researchers in 2013; Google has never documented it. It is not needed to run a search, and LinkClean strips it on Google's search domains.
- eiei is Google Search's session identifier. It is undocumented, but people who decoded it found that it begins with a Unix timestamp — the moment the search was run. A results link you forward therefore says when you searched. LinkClean strips it on Google's search domains.
- gs_lcrpgs_lcrp is added by Chrome, not by Google's results page: when you search from the address bar, Chrome encodes statistics about how you used the search box and sends them along as
gs_lcrp=. It is visible in Chromium's own source code. LinkClean strips it on Google's search domains. - rlzrlz is a promotional tag Chrome adds to Google searches made from the address bar. Unlike most of the search tail it is documented: Google says it is not unique to you, and that it records how Chrome was obtained, the week it was installed and the week of the first search. LinkClean strips it on Google's search domains.
Amazon
- dibdib is an opaque token Amazon attaches to a product link when you open it from a search results page, paired with
dib_tag(usuallyse). Amazon has not documented either. The product loads identically without them, and LinkClean strips both on Amazon storefronts — together withqid,sr,cridandsprefix, which restate the search you ran. - dchilddchild is a marker Amazon adds to product links, almost always as
dchild=1, typically on links opened from a search results page. Amazon has not documented it. The product is addressed by the ASIN in the path, and LinkClean stripsdchildon Amazon storefronts.
Email marketing
Mailchimp
- mc_cidmc_cid is Mailchimp's campaign identifier — names which Mailchimp campaign the link belongs to (the campaign-level identifier; mc_eid is the per-recipient one). Stripping mc_cid is less sensitive than mc_eid but pairs with it on every Mailchimp link.
- mc_eidmc_eid is Mailchimp's per-recipient identifier — a token tied to the specific email address the newsletter was sent to. Forwarding it tells Mailchimp that someone else opened your email. Of all the trackers LinkClean strips, this is the one that most directly leaks identity.
- mc_tcmc_tc is Mailchimp's tap-target identifier — names which specific link inside an email was clicked. The third in Mailchimp's standard tracking trio with mc_cid (campaign) and mc_eid (recipient). Strip it with the rest.
HubSpot
- _hsenc_hsenc is HubSpot's encoded engagement parameter — joins email clicks back to a specific contact record in HubSpot's CRM. Companion to _hsmi (the email's message ID). Same identity-bound risk as Mailchimp's mc_eid or Marketo's mkt_tok.
- _hsmi_hsmi is HubSpot's email message identifier — names the specific email send the link came from. Less sensitive than _hsenc (campaign-level vs per-recipient), but always paired with it on HubSpot links.
Social
Meta
- igshidigshid is the share-link identifier Instagram appends when you copy a link from the app — it ties the click back to your Instagram session for Meta's first-party analytics. The app has since renamed it twice, to
igshand thenigsi, and post links copied today carry a fourth name,stkn. LinkClean strips all four. Stripping preserves the destination perfectly while breaking the attribution chain back to your account. - mibextidmibextid is the share-identifier Facebook's mobile apps add to every link you share — separate from fbclid (which is for paid-ad clicks). It identifies the sharing session, not an ad click. Strip it before forwarding.
- ig_ridig_rid is an identifier Instagram attaches to some profile and post links, usually a value shaped like a UUID. Instagram has not documented it. The profile or post is addressed by the path, so the link works without it, and LinkClean strips it on instagram.com.
- fb_action_idsfb_action_ids is one of four tags that ride on a link Facebook opens from a story, with
fb_action_types,fb_sourceandfb_ref. Facebook's own SDK reads them as the IDs and types of the actions behind the link and a breadcrumb set when the story was made. The destination is in the rest of the address, and LinkClean strips all four on every site.
Region & language (preserved)
Google (Search & YouTube)
- hl
hlstands for “host language” — it tells Google services which language to render the interface in (hl=ja→ Japanese UI,hl=fr→ French). It's functional, not tracking. LinkClean preserves it on every host. The page is in this glossary because everyone asks what it is, not because we strip it. - gl
glstands for “geolocation” — it tells Google services (Search, Maps, Shopping) which country to scope results to.gl=usreturns US-relevant results;gl=jpreturns Japan-relevant results. Functional, not tracking. LinkClean preserves it. - t
t=on a YouTube URL means start playing at that many seconds in —?t=42jumps to the 42-second mark,?t=1m20sjumps to one minute twenty. Functional, not tracking. LinkClean preserves it host-scoped to youtube.com / youtu.be. - v
v=on a YouTube URL is the video identifier — the unique 11-character base64 ID that names the specific video. Strip it and the URL doesn't resolve to any video at all. Functional, not tracking. Always preserved.
Multi-vendor
- q
q=is the search-query parameter — a near-universal convention for naming the search term in a URL. Used by Google (google.com/search?q=…), Wikipedia, GitHub, Bing, DuckDuckGo, Stack Overflow, and many more. Functional, not tracking. LinkClean preserves it. - lang
lang=is the generic language parameter — a convention many non-Google sites adopt to select the interface or content language. Wikipedia usessetlang/uselang; many CMSes uselangdirectly. Functional, not tracking. LinkClean preserves it.
Clean tracking on iPhone, in one tap.
LinkClean strips 250+ vendor-specific tracking parameters from any link, from any app's share sheet — and preserves functional ones like hl, t (YouTube timestamp), and q (search). No account, on-device.
