llms.txt Adoption Report 2026
How many websites actually publish an llms.txt file for AI assistants? Measured on 1,408 real websites, not a survey.
Sample: 1,408 websites with a reachable homepage, out of 1,769 domains scanned · Scans from Sep 19, 2026 to Sep 22, 2026 · Recomputed Sep 25, 2026
33.6% of the 1,408 websites AIVIS could load publish an llms.txt that follows the llmstxt.org convention: a markdown file, opened by an H1 title, that tells AI assistants what the site is and where its key pages are. 45.5% return something at /llms.txt, but 26.1% of those responses have no markdown title at all — usually an HTML page served in place of a file that does not exist.
28.1% of all websites go further and link their key pages from the file, and 32.8% pass the full AIVIS llms.txt check (title plus sections or links). The median conformant file is 7,473 characters long.
Adoption depends heavily on the industry: among sectors with enough domains to be representative, SaaS & software leads with 57.8%, while Public Sector has the lowest rate at 3.7%.
llms.txt adoption by sector
Share of websites in each industry with a conformant llms.txt, and share that respond at /llms.txt at all.
| Sector | Websites | Conformant llms.txt | Respond at /llms.txt |
|---|---|---|---|
| All websites | 1,408 | 33.6% | 45.5% |
| SaaS & software | 128 | 57.8% | 63.3% |
| News & Media | 54 | 22.2% | 31.5% |
| Universities | 48 | 8.3% | 14.6% |
| E-commerce | 37 | 32.4% | 48.6% |
| Public Sector | 27 | 3.7% | 37% |
| Banking & Finance | 24 | 12.5% | 16.7% |
| Hotels & B&Bs | 13 | 23.1% | 23.1% |
| Restaurants | 11 | 9.1% | 27.3% |
| Telecom & Utilities(low sample) | 9 | 11.1% | 22.2% |
| Automotive(low sample) | 9 | 44.4% | 66.7% |
| Travel & Transport(low sample) | 6 | 16.7% | 66.7% |
| Real estate agencies(low sample) | 6 | 0% | 16.7% |
| Web agencies(low sample) | 4 | 75% | 75% |
| Law firms & professional services(low sample) | 2 | 50% | 50% |
| Local businesses & tradespeople(low sample) | 2 | 50% | 50% |
Sectors are assigned by the AIVIS classifier from homepage content: 390 of the 1,769 domains have one, the rest count in the totals but not in this table. Sectors with fewer than 10 websites are shown greyed out and marked low_sample in the dataset: too few to represent the industry.
Methodology
- Sample: the latest scan of every domain in the AIVIS corpus (1,769 domains, scanned between Sep 19, 2026 and Sep 22, 2026). Only the 1,408 domains whose homepage could be downloaded are counted: when a site refuses the crawler entirely, a missing llms.txt would say nothing about the site.
- Measure: the crawler requests /llms.txt at the root of each domain. A file counts as conformant when it contains a markdown H1 title, the one element the llmstxt.org proposal requires. Links and sections are counted separately; the check does not judge how well the content is written.
- Limits: “Respond at /llms.txt” is an upper bound — some sites answer any unknown URL with a normal web page and a 200 status, which is why the headline uses conformant files only. Only /llms.txt is measured, not /llms-full.txt or files in subfolders.
- No per-platform breakdown: the scanner does not record which CMS or website builder a site runs on, and this report does not estimate it. For online stores by platform, see the Agent Commerce Index.
- Relation to the AI Readiness Index: the Index reports the share of all scanned domains that pass the AIVIS llms.txt check; this report uses reachable homepages as its base and separates presence from conformance, so the two figures differ by definition. Both are recomputed live from the same scans.
Download the dataset
Totals and every sector row behind this report, recomputed live as the corpus grows.
Licensed under CC-BY 4.0 — free to reuse with attribution to AIVIS.
Cite as: AIVIS, llms.txt Adoption Report 2026, https://aivis.lumnika.com/en/research/llms-txt-adoption-2026
Does your site have an llms.txt?
Scan any domain for free: AIVIS checks the file, its structure and the seven other signals AI assistants rely on.