# Covering llms.txt

llms.txt covers the URL, at the root or its path

**In short:** Publish your llms.txt at https://your-domain/llms.txt on the production host (or at a section’s path, such as /docs/llms.txt), answering 200 to anyone.

- Category: Retrievability & Access · weight 30 of 100 within the category
- Retrievability & Access is worth 15 in every profile (points of 100, by scoring profile)
- Audit ID R1 (as shown on reports) · rubric v2.1

## What it checks

Whether an llms.txt file covers the URL you scanned, at the root (/llms.txt) or at its path (/docs/llms.txt).

## Why it matters

The llms.txt proposal puts the file at a fixed, predictable URL, the root or the path of the pages it covers, so an agent or tool that goes looking for it knows where to ask. If no file covers the page, nothing else in the file matters.

## When it applies

Always, for any llms.txt that exists. A site with no covering file isn’t scored on R1; it’s reported as Not Present.

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

## How it’s scored

1.0 for the most specific file covering the scanned URL, root or path. With no covering file at the URL’s path or any ancestor path, the scan reports Not Present (the no-file cap of 0) instead of a score.

> **Measurement.** This audit is not yet measured precisely; everyone who qualifies gets the same score.

> **llms.txt v2.** llms.txt v2 (Aug 2026) defines subpath files: /docs/llms.txt covers everything under /docs/, and agents should use the most specific file. Rubric v2.0 gives full credit for the most specific covering file, root or path (v1.1 gave 0.5 for a subpath file), and the report names the file that covers your URL.

## How to fix

1. Publish the file at https://your-domain/llms.txt on the production host, not only on staging.
2. If one section of your site has its own audience (docs, developers), you can publish /docs/llms.txt too; it covers everything under /docs/.
3. Check it returns 200 from outside your network, with no login or IP allowlist in front of it.

## Frequently asked questions

### Does a /docs/llms.txt count if I don’t have one at the root of my site?

Yes. Under llms.txt v2 a file covers the URLs under its path, so a scan of a page under `/docs/` is fully covered by `/docs/llms.txt` and earns full marks with no root file. A scan of your bare domain only looks for `/llms.txt`, though, so publish a root file too if you want the homepage covered.

Sources: [llms.txt (v2, Aug 2026)](https://llmstxt.org/) · [Covering llms.txt](https://llmsrank.dev/docs/audits/covering-llms-txt.md)

### I have both /llms.txt and /docs/llms.txt. Which one do you score?

The most specific file that covers the URL you scanned, as llms.txt v2 tells agents to do. We try each ancestor directory from the deepest up (`/docs/api/llms.txt`, then `/docs/llms.txt`, then `/llms.txt`, at most five tries with the root always last) and the first that answers 200 with a real text file wins; the report names it as the covering file. A last path segment without a file extension is also tried as a directory, so a scan of `https://example.com/docs` (with or without the trailing slash) starts at `/docs/llms.txt`, while `https://example.com/docs/page.html` is a page and starts at its parent directory.

Sources: [llms.txt (v2, Aug 2026)](https://llmstxt.org/) · [Covering llms.txt](https://llmsrank.dev/docs/audits/covering-llms-txt.md)

### Does my llms.txt on docs.example.com cover www.example.com, or the other way round?

No. Coverage runs down paths within one site: we only look for covering files on the host you scanned, so each subdomain needs its own `/llms.txt`. If your main file lives on a subdomain, publish a short root file on the main domain as well that links to it.

Sources: [llms.txt (v2, Aug 2026)](https://llmstxt.org/) · [Covering llms.txt](https://llmsrank.dev/docs/audits/covering-llms-txt.md)

### Does llms-full.txt, or an HTML page at /llms.txt, count as a covering llms.txt?

Neither does. We look only for files named `llms.txt`; `llms-full.txt` is captured separately for research and never stands in for it. A path that answers 200 with an HTML page (by body or by a `text/html` content type), or with a body under 20 bytes, isn’t treated as a file: resolution moves on to the next ancestor, and if nothing else covers the URL the scan reports that a web page answered instead.

Sources: [llms.txt (v2, Aug 2026)](https://llmstxt.org/) · [Covering llms.txt](https://llmsrank.dev/docs/audits/covering-llms-txt.md)

### In your research, how many sites earn full marks for having a covering llms.txt?

All 156 tracked sites score 1.0 here, but that is not a finding: the corpus is made of sites that publish an llms.txt, and any covering file earns full credit by design, so this audit is not yet measured precisely. Sites without a covering file are reported as Not present instead of being scored.

Sources: [LLMSRank method](https://llmsrank.dev/docs/method.md) · [Covering llms.txt](https://llmsrank.dev/docs/audits/covering-llms-txt.md)

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