Every URL in a creator's bio is recorded and normalized into a consistent format, with link-in-bio services flagged (Linktree, Beacons, etc.). Only links seen within the last 12 months are kept, anything older is dropped. This tells you where a creator is active outside the platform right now, not where they used to be.
Share of profiles where we found at least one link in the bio.
Four stages take raw bio text to a clean, deduplicated list of links with link-in-bio classification.
Pattern matching finds every URL-like string in the bio text, including partial URLs without a protocol.
Strip zero-width characters, trailing punctuation, and artifacts. Lowercase domains, convert mobile URLs, unify to HTTPS.
Combine URLs found in bio text with any links the platform itself provides (e.g. the website field). Remove duplicates.
Match each normalized URL against a list of known link-in-bio domains. Flag the profile if any match is found.
Each link is checked against a maintained list of known link-in-bio services. We review which services creators are using every month, so new platforms get added as they gain traction.
Creators who use a link-in-bio service typically manage multiple destinations: a shop, a booking page, other social profiles, or content hubs. This is a strong signal that they are treating their online presence as a coordinated operation rather than a single-platform activity. For brands, this can indicate a creator who is more commercially active and easier to reach through multiple channels.
Personal site, link aggregator, and cross-platform link
Only a personal blog, no link aggregator present
NikolaTravels has three distinct links including a Linktree page, showing an active multi-channel presence. A brand searching for creators with organized link setups would see this profile as commercially ready. BorisLifeLog has a single personal blog link and no link aggregator, suggesting a simpler setup focused on content rather than multi-channel distribution.
Since this is a pattern-based pipeline (not a subjective classification), validation focuses on two things: how completely it finds URLs, and how accurately it normalizes them.