llms.txt Checker & Generator
Validate your site's llms.txt against the llmstxt.org format — and if it's missing or weak, generate a ready-to-paste starter file built from your own pages. Works on any site, including SPAs, because we read the literal file, not the rendered page.
What we check for you
19 concrete checks across 6 sections — graded against the llmstxt.org format, plus a deterministic starter-file generator.
File Availability
Reachable at /llms.txt, served as real text (not an HTML/SPA fallback or soft-404), non-empty, and valid UTF-8 with a BOM tolerated.
5 checks
Required Structure
A single H1 site/project name — per llmstxt.org the only mandated element. We also flag multiple H1s and sentence-like names.
3 checks
Summary & Intro
The optional one-line blockquote summary after the H1, and a heading-free intro region (the spec allows any content there except headings).
2 checks
Sections & Links
H2 link sections where every item is a [name](url) link, the exact ": " description separator, absolute/root-relative targets, and the special ## Optional section.
5 checks
Link Validity (sampled)
A bounded live sample (up to 10 deduped links) of HEAD requests — broken 4xx/5xx links and links that resolve only via redirects. Not an exhaustive crawl.
2 checks
llms-full.txt Bonus
Detects the optional /llms-full.txt convention file (a Mintlify-style tooling convention, not part of the canonical spec) — a bonus, never required.
2 checks
Plus: a ready-to-paste starter file
When your file is missing or weak, we build a correctly-formatted llms.txt from your homepage navigation and sitemap.xml — assembled deterministically, not written by AI. Review the names, URLs, and descriptions before you publish it.
What is llms.txt?
A Markdown file at your site root that points AI agents to your most important pages. It is a proposed community standard from Jeremy Howard / Answer.AI (Sept 2024) — not an official W3C or IETF standard — and adoption by AI crawlers is not guaranteed. Source: llmstxt.org.
We fetch the literal /llms.txt
An SSRF-guarded request reads the raw file at your site root — the actual bytes, never the rendered DOM. That is why it works even on client-rendered SPAs.
We validate structure & links
We parse the Markdown and grade it against the llmstxt.org format — the required H1, summary, H2 link sections, link formatting — and live-sample a few links.
We generate a starter (if needed)
If the file is missing or weak, we assemble a correctly-formatted starter from your homepage navigation and sitemap.xml — deterministically, with no AI.
We grade the file, not the outcome. A present, well-formed llms.txt is a clean signal to AI agents — but no file guarantees an AI will read, honour, or cite your content. We report “file present & well-formed”, never “AI-optimised” or “guaranteed indexed”.
Checking it is step one. Keeping it current is the goal.
Ampifex generates and maintains your llms.txt as you publish — and cross-links into the broader AI-visibility program.
Generated From Your Pages
Ampifex builds your llms.txt from the collections it already publishes — real names and URLs, kept in the llmstxt.org format.
Kept Current Automatically
Every time you publish, your llms.txt is regenerated — no stale links, no manual edits, no drift from your live content.
Well-Formed By Default
A single H1, a clean summary, properly-formatted [name](url): note sections — the structure this checker grades, handled for you.
Part of Your AI-Visibility Program
llms.txt is one signal. Ampifex folds it into the broader GEO / AI-visibility work so your content is readable end to end.
Keep your llms.txt current — automatically.
Ampifex already manages your content publishing, so keeping llms.txt in sync with what you publish is a genuine deliverable, not a vanity promise.