Clioos › Guides › llms txt Example
llms txt Example
Most llms.txt examples online are hand-typed illustrations. This one is not: it is the exact file the free Clioos generator produced on 2 November 2026 for a fictional example site, together with the checker results for that file and for a deliberately broken one. Copy the shape, not the example's content — replace the bakery with your own site, keep the structure, and run the same checks before publishing.
At a glance
- The example below is verbatim output from POST /llms-txt/api/generate for a fictional “Example Bakery” site — not a hand-typed illustration.
- Structure in order: H1 title, blockquote summary, then ## sections of - [name](url): notes links.
- The generator auto-added the homepage link to the Docs/Pages section; the two links supplied (Menu, Order) kept their notes.
- Checker verdict on this file: looks_like_llms_txt true, has_h1 true, 1 section, 3 links, valid true, zero warnings and zero errors.
- Checker verdict on plain text with no H1: valid false, with the error “Missing required H1 title (for example, '# My Site').”
- Generation is pure templating (no LLM call, no crawl). The checker is structural only — it cannot judge whether your summary or link choices are good.
Steps
- Read the example file in the next section line by line. Every line is real generator output; the only fictional parts are the bakery name, summary, and example URLs (example-bakery.test is a placeholder domain).
- Replace the values with your own: your H1 is your site or project name, your blockquote is one honest summary sentence, and your links are the pages you most want a reader — human or model — to find first.
- Add more ## sections when you have distinct groups (for example Docs, Products, Guides). Keep each link on one line in the form - [name](url): notes, with a short note on what the page is.
- Optionally add a ## Optional section for secondary links, and optionally sketch an llms-full.txt outline for tools that want more detail. Both are optional; neither is required for a valid file.
- Check the finished file with the Clioos checker (paste the body, or point it at the public URL once published). Publish only when it reports an H1 present and no errors, and verify https://yoursite.com/llms.txt returns Markdown, not HTML.
The example file
Generated 2 November 2026 from this input: site name “Example Bakery”, summary “A small bakery site with menus and ordering pages.”, homepage https://example-bakery.test/, important links Menu (https://example-bakery.test/menu, note “Current menu”) and Order (https://example-bakery.test/order, note “Online ordering”). The output, verbatim:
# Example Bakery
> A small bakery site with menus and ordering pages.
## Docs/Pages
- [Homepage](https://example-bakery.test/)
- [Menu](https://example-bakery.test/menu): Current menu
- [Order](https://example-bakery.test/order): Online ordering
Line by line: the H1 is the site name and is required — without it the file fails the check (see below). The blockquote line is the optional summary. The ## Docs/Pages heading starts the first link section; the generator groups the homepage URL and important links here. Each link line is a Markdown link plus, where supplied, a colon and a short note. The homepage link has no note because none was supplied — that is fine, notes are optional.
Worked example (generate + check, tested 2 November 2026)
Test method, disclosed honestly: both calls hit the live llms-txt service's own API on the server (POST /llms-txt/api/generate, then POST /llms-txt/api/check with the generated body, plus a second check with plain text that has no H1). No account, no payment, and no LLM call were involved, and the no-code version of the same job is the form at https://clioos.com/llms-txt/ .
| Stage | Result |
|---|---|
| Generate | HTTP 200 in 0.03 seconds; validation valid: true, has_h1: true, h1_count: 1, section_count: 1, link_count: 3, errors: [], warnings: [] |
| Check, generated file | HTTP 200; looks_like_llms_txt: true, has_h1: true, section_count: 1, link_count: 3, valid: true, warnings: [], errors: [] |
| Check, plain text without H1 | HTTP 200; looks_like_llms_txt: false, has_h1: false, section_count: 0, link_count: 0, valid: false; warning and error: “Missing required H1 title (for example, '# My Site').” |
Two things are worth noting. First, the generate and check calls agree: the same file that generates clean also checks clean, which is what a templating-plus-structural-check pipeline should do. Second, the broken input actually fails — a checker that passed everything would be decoration, and this one is not.
Checklist before you publish
- Exactly one H1, at the top: the file must open with a single “# Your Site” title. Missing H1 is the failure the checker reports most plainly.
- Summary is honest: one blockquote sentence describing what the site is. The checker cannot verify honesty — you must.
- Every link resolves: click each URL. A structurally valid file pointing at dead links is still a bad file.
- Link notes are informative: “Current menu” tells a reader something; “Click here” tells them nothing. Write notes for humans first.
- Sections are grouped by purpose: do not dump fifty links under one heading. Split by Docs, Products, Guides, or whatever your site's real groups are.
- The root URL returns Markdown: fetch https://yoursite.com/llms.txt after publishing. A 404, a homepage redirect, or an HTML page means the file is not published, whatever your editor says.
- Optional sections stay optional: a ## Optional section and an llms-full.txt outline are welcome extras, not requirements. Skip them rather than padding them with junk links.
What it can't do
- It cannot make AI tools cite you: publishing llms.txt does not guarantee citations, rankings, or traffic. The example above is valid and still promises nothing.
- The checker cannot verify meaning: it reports structure (H1 present, section count, warnings). It cannot tell whether your summary is accurate or your link choices are wise.
- The generator cannot discover your pages: it templates what you type and does not crawl your site. Forgotten pages stay forgotten.
- It cannot fix a bad site map: if your important pages are buried, thin, or dead, listing them in llms.txt advertises the problem rather than solving it.
- It is not a standard to appeal to: llms.txt is a proposal. If a tool ignores your file, there is no compliance process — only the option to keep the file anyway, since it costs almost nothing.
Frequently asked questions
Can I copy this example file directly?
Copy its structure, not its content. Replace the bakery name, summary, and example-bakery.test URLs with your own site's values, then check the result before publishing.
What makes an llms.txt file valid?
Structurally: an H1 title at the top, plus ## sections of - [name](url): notes links. The Clioos checker reports H1 presence, section count, and structural warnings, and rejects bodies with no H1.
What is the ## Optional section for?
Secondary links that are useful but not essential. It is optional — a file without it is still valid, and padding it with junk links helps nobody.
What is llms-full.txt?
An optional fuller companion outline for tools that want more detail than the short llms.txt map. The Clioos generator can sketch one from the same input; it is optional, not required.
Does this example guarantee anything about AI citations?
No. The file checks as valid and still guarantees no citations, rankings, or traffic. Treat llms.txt as documentation with a possible upside, not a ranking lever.
Related guides: What Is llms txt · llms.txt Generator & Checker tool · The llms.txt proposal (llmstxt.org) · All guides · Tool: PDF Probe