# URL hygiene

URL hygiene — absolute, canonical, no params

**In short:** Link each page once, by its full https:// URL, with no tracking parameters and no redirects.

- Category: Curation Quality · weight 12 of 100 within the category
- Curation Quality is worth 25 in generic and commerce, 20 in regulated and technical (points of 100, by scoring profile)
- Audit ID C5 (as shown on reports) · rubric v2.1

## What it checks

Whether link URLs are absolute, canonical, parameter-free and redirect-free.

## Why it matters

Relative, tracking-tagged or redirecting URLs make agents fetch the wrong page, or the same page twice, and waste their budget.

## When it applies

Applies when the file has links. A file with no links is Not applicable here.

An audit that doesn’t apply, or that we couldn’t verify, is removed from your score and its weight shared among the Curation Quality audits that do apply. It is never scored as a zero.

## How it’s scored

0.6 × the share of links that are absolute (start with http:// or https://) + 0.4 × the share that are unique (compared as exact strings).

> **Measurement.** Today only two of the four sub-checks are measured. Query parameters and redirects aren’t checked, and uniqueness compares exact strings, so /pricing and /pricing/ count as two URLs. Checking canonical, parameter-free and redirect-free URLs is planned.

## How to fix

1. Use full https:// URLs, not relative paths.
2. Strip query strings and tracking parameters (utm_*, ref).
3. Link to the final, canonical URL rather than one that redirects.
4. List each URL once.

## Frequently asked questions

### How is URL hygiene calculated?

Today it is `0.6 × share of absolute links + 0.4 × share of unique links`. A link is absolute when it starts with `http://` or `https://`, and uniqueness compares the URL strings exactly. So a file whose links are all relative scores 0.4 at most, and one with 10 absolute links to 8 distinct URLs scores 0.92.

Sources: [LLMSRank — URL hygiene audit](https://llmsrank.dev/docs/audits/url-hygiene.md)

### Does linking the same page twice hurt?

Yes, twice over: each repeat lowers the unique share here (10 links to 8 distinct URLs is 80% unique, costing 0.08) and adds to the link count that [Signal-to-noise](https://llmsrank.dev/docs/audits/signal-to-noise.md) measures. Because the comparison is by exact string, `/pricing` and `/pricing/` aren’t caught as a repeat, but they are still one page listed twice for an agent. Pick one form of each URL and list it once. The [consistency checker](https://llmsrank.dev/tools/consistency) lists each mismatched URL form with the exact line to fix.

Sources: [LLMSRank — URL hygiene audit](https://llmsrank.dev/docs/audits/url-hygiene.md) · [Google Search Central — Consolidate duplicate URLs](https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls)

### Should I remove query strings and UTM parameters from my links?

Yes. Tracking parameters such as `utm_source` or `ref` don’t change the page, so they only split one page across several URLs; Google’s advice is to trim “parameters that don’t change the content”. Today’s computed score doesn’t parse query strings (a tagged URL still counts as absolute and unique), but the rubric asks for parameter-free links, so strip them anyway.

Sources: [Google Search Central — URL structure best practices](https://developers.google.com/search/docs/crawling-indexing/url-structure) · [LLMSRank — URL hygiene audit](https://llmsrank.dev/docs/audits/url-hygiene.md)

### Are relative, protocol-relative or http:// links acceptable?

Relative paths such as `/docs/refunds.md` and protocol-relative ones such as `//acme.com/docs/refunds.md` are both relative references under RFC 3986, so neither counts as absolute and together they can cost up to 0.6. An `http://` link does count as absolute, but link the final `https://` URL anyway so an agent isn’t sent through a redirect: the rubric asks for canonical, redirect-free links.

Sources: [RFC 3986 — URI Generic Syntax, §4.2 Relative Reference](https://datatracker.ietf.org/doc/html/rfc3986#section-4.2) · [LLMSRank — URL hygiene audit](https://llmsrank.dev/docs/audits/url-hygiene.md)

### How do real llms.txt files do on URL hygiene?

In our research, 174 of the 187 sites in our corpus (captured 2026-10-10, scored with rubric v2.1) had links to score, and 117 (67.2%) earned full marks; the mean was 0.967, the highest in the Curation category. netlify.com, expedia.com, tavily.com, you.com and cloudflare.com are examples.

Sources: [LLMSRank — Method](https://llmsrank.dev/docs/method.md) · [LLMSRank — URL hygiene audit](https://llmsrank.dev/docs/audits/url-hygiene.md)

See also: [all audits](https://llmsrank.dev/docs/audits.md) · [how scoring works](https://llmsrank.dev/docs/method.md)