What llms.txt Does and Does Not Do

llms.txt is a small, dated file at the root of your site that gives language models a curated index of what you sell and where the answers live. It helps engines read your site faster and quote it more accurately; it does not force any engine to cite you, and it is not a ranking trick.

What it does (with our file as the live example — /llms.txt, shipped and verified 11 Oct 2026)

• A one-paragraph brand statement: who you are, legal entity, registration number, founder, location — the entity block engines cross-check.
• A products section with real names, real prices, real access terms (G7: our file carries Scan US$149/RM640, Scale US$199/RM850, the advisory ladder from US$3,000).
• Method and company links, so a model can find the rubric and the policies without crawling.
• A "Last updated" date, so staleness is checkable.

What it does not do (the honest half)

1. It is not a gate. Engines do not require it; blocking crawlers in robots.txt does far more damage than a missing llms.txt.
2. It is not a ranking signal in the SEO sense. There is no llms.txt score in a search index; it is a reading aid for models that look for it.
3. It cannot override your pages. If the site says one thing and llms.txt another, engines trust the pages; keep them consistent and bump the date when they change.
4. It does not substitute for schema. Structured Product/Offer markup is what lets an engine quote price and currency without guessing.
5. It guarantees nothing. Engines choose what to cite; the file removes a reading excuse, not the outcome (skillpilotadvisory.ai/geo FAQ: no citation guarantees, fetched 10 Oct 2026).

The 10-minute version (copy-paste skeleton)

# {Brand}
> {One sentence: what you sell, for whom, where you are}
## Products
- [{Product}]({URL}): {what it includes}, {real price}, {access period}
## Company
- [About]({URL}) | [Contact]({URL}) | [Terms]({URL}) | [Privacy]({URL})
Last updated: {date}

Update it whenever prices or products change — an outdated index is worse than none (it teaches engines your file lies).

Sources and context

• Our own file follows this structure and is linked from the live GEO pillar (/pillar/become-the-answer links /llms.txt, live 11 Oct 2026).
• "Most small and mid-size sites are effectively invisible to AI crawlers. No llms.txt, thin entity markup, and crawler rules written for search, not for answers" (skillpilotadvisory.ai/geo, fetched 10 Oct 2026).
• The discovery-files point is one of seven in the 40-check audit method (checks 27-31, /pillar/become-the-answer, live 11 Oct 2026).

FAQ

Is llms.txt official? It is a community proposal adopted by sites as a reading aid; engines may use it, and none require it.
Where does the file live? Site root, /llms.txt, plain text.
How often to update? Every price or product change; the date line is the proof.
Does it replace a sitemap? No — sitemap for crawlers, llms.txt for models; keep both.