AI Citation URL Normalizer and Publisher Concentration Check

1. Paste observed citation URLs

One URL per line. Extra text is allowed when the line contains one complete http or https URL.

Citation observations

The example domains are reserved placeholders and do not represent real AI-answer evidence.

Remove common analytics parameters such as utm_*, gclid, fbclid, msclkid, mc_cid, and _hs*. Also remove ref, referral, affiliate, and affiliate_id. Enable only when these are not content identifiers. Remove a trailing slash from non-root paths for comparison. Sort retained query parameters by key and value.

2. Review the comparison ledger

Duplicates are calculated from normalized keys, not from live canonical or redirect checks.

0valid observations

0unique URL keys

0unique publishers

0duplicate variants

Run the synthetic example to calculate publisher concentration.

No observations analyzed yet.

Methodology and audit boundaries

AI-answer citation exports often contain tracking parameters, fragments, repeated paths, or multiple observations from the same publisher. A stable local comparison key reduces accidental double-counting while keeping the original URL beside every transformation. The companion citation URL normalization protocol documents the rules and audit boundaries.

1 · Preserve raw input

Keep the exact observed URL and a dated answer snapshot outside this tool. The export includes the original string for auditability.

2 · Normalize conservatively

Lowercase the hostname, remove fragments, remove selected tracking parameters, and preserve path/query meaning.

3 · Separate counts

Report observations, unique URL keys, and unique publishers as different measures. They answer different questions.

4 · Verify manually

For consequential analysis, follow redirects and inspect publisher canonicals, content equivalence, dates, and source roles.

Transformation	Default	Reason and caveat
Lowercase hostname	Always	DNS hostnames are case-insensitive. Path casing remains untouched because it may be meaningful.
Remove URL fragment	Always	Fragments are not sent to the server, but they can still identify a section; the original URL remains in the ledger.
Remove analytics parameters	On	Known click/analytics IDs commonly create duplicate variants. Unknown parameters are retained.
Remove referral parameters	Off	Referral fields can be meaningful to attribution or routing, so removal requires an explicit choice.
Change scheme or resolve redirect	Never	HTTP and HTTPS, redirects, and canonical targets require network verification; this client-side utility does not guess.