Paste your llms.txt below and I’ll check it against the official Answer.AI spec in seconds — line-by-line errors, a compliance score out of 100, and an auto-corrected version you can copy straight back to your server.
Open yoursite.com/llms.txt in a new tab, select all (Ctrl/Cmd+A), copy, and paste it here. Everything runs in your browser — nothing is uploaded anywhere. (Browsers block cross-site fetching for security — CORS — so paste or upload is the reliable path.)
Waiting for input…
Paste your llms.txt above to see your compliance score.
Spec checklist
- iNo content yetPaste your file to run the 11 spec checks.
Note: live HTTP link checking (does each URL actually return 200?) is not possible client-side. This tool validates URL format and structure; spot-check a few links manually or crawl them server-side.
How does your file compare? (benchmarks, est.)
| Metric | Well-formed llms.txt (est.) | Your file |
|---|---|---|
| Compliance score | 90–100 / 100 | – |
| File size | 1–20 KB | – |
| H2 sections | 2–8 | – |
| Link entries | 8–60 | – |
| Entries with descriptions | 90%+ | – |
Why I built a free llms.txt validator
I run Sprout Sage Solutions, and a growing share of the traffic my clients care about no longer comes from a blue Google link — it comes from ChatGPT, Perplexity, and Claude answering a question and citing (or not citing) their site. The llms.txt file, proposed by Answer.AI, is the closest thing we have to a sitemap for AI assistants: one small markdown file at your root that tells an AI which pages matter and what each one covers.
The problem I kept running into on audits: most llms.txt files in the wild are broken. Not slightly — structurally. Relative URLs an AI can’t resolve. Full page dumps pasted where a short index should be (that’s llms-full.txt’s job). Two H1s. Sections with no links. Plain text with no markdown at all. Each of those is a page an AI assistant quietly skips, and you’ll never see it in any report. In my own site reviews, I’d estimate well over half of published llms.txt files fail at least one spec check (est. — based on files I’ve audited, not a formal study).
How the validator works
Everything runs in your browser — your file is never uploaded to my server or anyone else’s. Because browsers block cross-site fetching (CORS), the tool can’t pull the file from your URL directly; open yoursite.com/llms.txt, copy the contents, and paste them in, or upload the file. The script then runs 11 checks against the spec: exactly one H1 as the first line, the optional blockquote summary, H2-delimited sections, [title](url): description link syntax, absolute https URLs, duplicates, empty sections, the “Optional” naming convention, file size (I flag anything over ~50 KB, est. practical limit), and heuristics that catch plain-text files and llms-full.txt-style content dumps. You get line numbers, the offending line, a score out of 100, and an auto-corrected version where the fix is mechanical.
If your score is low, the fastest path is usually to regenerate rather than patch. And if you want the bigger picture — whether AI assistants can find, crawl, and cite your whole site — that’s exactly what my answer engine optimization service covers. For med spa owners specifically, I’ve written up how AI search is changing patient acquisition in my med spa marketing guide, and while you’re here, my missed call calculator shows what the leads you already generate are worth when nobody answers the phone.
Frequently asked questions
What is llms.txt supposed to contain?
A markdown index: one H1 with your site name, an optional one-line blockquote summary, and H2 sections (like “## Services” or “## Docs”) each containing a bullet list of [page title](https://full-url): short description entries. It’s a curated menu for AI assistants, not a copy of your content.
What’s the difference between llms.txt and llms-full.txt?
llms.txt is the short index of links. llms-full.txt is a separate, optional file containing the full expanded content for models that want everything in one request. The most common mistake I see is pasting entire pages into llms.txt — this validator flags that pattern specifically.
Why can’t the tool just fetch my llms.txt from my URL?
Browser security (CORS) blocks one website’s JavaScript from reading files on another domain, so a purely client-side tool can’t fetch your file. That’s also why your data stays private here — nothing you paste ever leaves your browser. For the same reason, the tool validates URL format but can’t live-check whether each link returns a working page.
Does Google or ChatGPT actually read llms.txt?
Adoption is still early and uneven. Some AI crawlers and tools read it today, others don’t yet, and nobody publishes firm numbers (any figure you see is est. at best). My take: it costs you ten minutes, it can only help, and a broken file is worse than none because it signals sloppiness on the exact pages you want cited.
What does the “Optional” section do?
The spec gives the literal section name “Optional” special meaning: AI agents can skip those links when they’re short on context window. Put nice-to-have pages (blog archives, changelogs) there and keep your money pages in the main sections.
Want me to run these numbers with you? Book a free strategy call or call/text me at +91 97297 12388.
