4. The plain-text summary for machines
What this is
A short file at the root of the site that tells an assistant, in plain markdown, what the business is and which pages matter.
Why this matters for auditlodge.com
An assistant that lands on auditlodge.com has to work out what the business is from the HTML. This file tells it directly, in 12 lines built from your own pages.
What we found on auditlodge.com
Built from 7 of this site’s own pages, using their real titles and descriptions - we wrote no sentence of it
We built the file below from 7 pages of auditlodge.com — your own titles and your own descriptions. We wrote none of its sentences.
- 3 list items under ‘When to use Auditlodge’ have no
[name](url)link, which their spec requires - no rel=“describedby” link pointing at the llms.txt - v2 of their spec added this and almost no tool emits it yet
- no rel=“alternate” type=“text/markdown” link to a markdown twin
Notes from reading your sitemap: robots.txt declared 1 sitemap; https://auditlodge.com/sitemap.xml -> 7 urls.
Step 1 — Publish the file at your root
Save this as llms.txt so it answers at https://auditlodge.com/llms.txt:
# Auditlodge — find out if anyone can actually find you
> Free check first — we read up to 1,000 pages of your site and show what stops people finding you. The full audit is from £9.99, paid once — the price follows the size of your site.
## Pages
- [Auditlodge — find out if anyone can actually find you](https://auditlodge.com/): Free check first — we read up to 1,000 pages of your site and show what stops people finding you. The full audit is from £9.99, paid once — the price follows the size of your site.
- [Fix your site — your audit, ready to apply | Auditlodge](https://auditlodge.com/fix-it): Your audit, ready to apply: the report, the copy-paste fixes file, the spreadsheet for your web person, and the guide. One payment, done in an afternoon.
- [Welcome to the Lodge - The Lodge Guide](https://auditlodge.com/learn/): People go to the lodge before they go on the trip. They get warm, they get a map, they get everything they need to succeed — and then they climb.
- [Auditlodge — your free check](https://auditlodge.com/result): A one-off website audit. We check whether anyone can find you, and we re-check every finding against your live site before we send it. From £9.99, paid once — the price follows the size of your site.
- [Privacy · Auditlodge](https://auditlodge.com/static/privacy.html): How Auditlodge handles your data: one email address to send your report, no account, no tracking of your visitors, and the registered company behind it.
- [Terms · Auditlodge](https://auditlodge.com/static/terms.html): The terms of the Auditlodge website audit: one payment that follows the size of your site, the full report pack, and no subscription.
- [What will you get? | Auditlodge](https://auditlodge.com/what-you-get): A real audit shown in full: what went wrong, what we fixed the same day, and the dated final audit that prints the last word.
Step 2 — Publish the 7 markdown copies
Each of your pages gets a plain-markdown twin at its own address. They are written for you already and sit in packs/04-llms-txt/markdown/. The addresses they belong at:
- https://auditlodge.com/fix-it.md
- https://auditlodge.com/index.html.md
- https://auditlodge.com/learn/index.html.md
- https://auditlodge.com/result.md
- https://auditlodge.com/static/privacy.html.md
- https://auditlodge.com/static/terms.html.md
- https://auditlodge.com/what-you-get.md
Step 3 — Point at them, one way or the other
Use either form. Not both.
As HTML, in the <head>:
<link rel="describedby" href="https://auditlodge.com/llms.txt">
<link rel="alternate" type="text/markdown" href="https://auditlodge.com/index.html.md">
Or as an HTTP response header. It also covers the markdown files themselves, and you set it in your server or CDN without touching a page:
Link: <https://auditlodge.com/index.html.md>; rel="alternate"; type="text/markdown", <https://auditlodge.com/llms.txt>; rel="describedby"
In the words of the company that published this
llmstxt.org: “The file can be placed at the site root, or at any path within it, covering the pages under that path.” The markdown copies go “at the same URL as the original page, either with .md appended (page.html.md) or with the extension replaced by .md (page.md)”. The pointer “can be provided as HTML <link> elements, or as an HTTP Link: response header… and can be added in web server or CDN configuration without modifying any pages.”
Their page: llmstxt.org
How to check you did it right
- Open https://auditlodge.com/llms.txt in a browser. It should return the file, not a 404, and start with
# Auditlodge — find out if anyone can actually find you. - Open one of the markdown addresses above; it should return plain text, not HTML.
What it looks like when it is working
- https://auditlodge.com/llms.txt answers, the pointer is in place, and the faults listed above stop being reported.
- The faults to clear: 3 list items under ‘When to use Auditlodge’ have no
[name](url)link, which their spec requires; no rel=“describedby” link pointing at the llms.txt - v2 of their spec added this and almost no tool emits it yet; no rel=“alternate” type=“text/markdown” link to a markdown twin.
In two lines
- An assistant that lands on auditlodge.com has to work out what the business is from the HTML. This file tells it directly, in 12 lines built from your own pages.
- First publish the file at your root, then publish the 7 markdown copies, then point at them, one way or the other.
The words in plain English
<head>— the hidden top part of a web page’s code, where settings and tags live - visitors never see it, machines read it first.- robots.txt — a small public text file at yoursite.com/robots.txt that tells crawlers which parts of the site they may read.
- sitemap — the list of your pages, at yoursite.com/sitemap.xml on most sites, that search engines treat as your site’s own inventory.
- llms.txt — a plain-text summary of your site, placed at the root, written for the AI assistants to read.
- markdown — plain text with light formatting marks; a markdown twin is a plain-text copy of a page at its own address.
- HTTP header — a line your server sends along with a page - invisible to visitors, read by machines.
- CDN — the delivery service that hands out your pages from servers near the visitor; many sites set their headers there.
- answer engine — an AI assistant - ChatGPT, Claude, Perplexity, Google’s AI - that answers a question by quoting pages it can read.
The finished file is in your pack at packs/04-llms-txt/llms.txt.