llms.txt Valid Link Rate reports the share of parsed Markdown link candidates that resolve to valid HTTP or HTTPS destinations. A file with no candidates returns 0. Duplicates can lower the rate, because candidate correctness and destination breadth are deliberately kept separate.
| Factor ID | RG-AIX-037 |
|---|---|
| Family | AI access & extractability |
| Measurement | Percentage |
| Measured zone | Validated root llms.txt links |
What it measures
This factor is a hygiene check on the file’s link syntax. It answers how much of what looks like a link actually works as one, which is the difference between an index a machine can follow and a list that silently drops entries.
How RankGear measures it
RankGear divides the number of distinct valid destinations by the number of Markdown link candidates and multiplies by 100. A file with no candidates returns 0.
| If the value is… | Do this |
|---|---|
| Below 100 with candidates present | Look for malformed destinations, unsupported schemes or duplicated entries and correct the syntax. |
| At or near 100 | Read it as confirmation the link syntax is sound; no change needed. |
| 0 with content present | Check whether the file contains link candidates at all, or whether none of them resolve. |
How to optimize it
Correct malformed destinations, unsupported schemes, accidental duplicates and stale link syntax while preserving genuinely useful resources. The aim is that every line that looks like a link behaves like one.
Important considerations
- Duplicates can lower the rate because distinct destinations are compared against all candidates.
- Candidate correctness and destination breadth are intentionally kept as separate measurements.
- A file with no link candidates returns 0 rather than being excluded.
- This is a measured correlation within a result set, not a guarantee: the value relates to how the pages ranking for your query are built — it does not, on its own, make a page rank.
Related factors
Part of the Factors reference · how RankGear measures · glossary.