Learn · 2026-09-30 · 2 min read
What is llms.txt and should your site have one?
llms.txt is a plain-text, Markdown-formatted file served at the root of a website (yoursite.com/llms.txt) that gives AI crawlers and answer engines a concise, human-written summary of what the site is, the facts that matter about it, and links to its most useful pages. It follows the llmstxt.org convention. It is not a ranking signal and not every engine reads it, but it costs an hour, it is the one place your facts are stated in one voice, and an AI visibility tool that did not have one would be embarrassing. helloAISearch's is at /llms.txt.
Free 3-day trial · No credit card · Cancel anytime
What the convention asks for
- An H1 with the site or product name.
- A blockquote with a one-paragraph summary.
- Optional sections of key facts in plain sentences.
- One or more H2 sections of links, each as `- [Title](url): one-line description`.
- Plain text, UTF-8, served with a text/plain or text/markdown content type.
That is the whole specification. The value is in the discipline it forces: you have to write down, in one place and in plain words, what you are, who runs you, what you cost, and where the useful pages are.
See what the engines say about your brand.
Ten of your prompts on ChatGPT and Gemini, twice, with the actual answers and your brand highlighted.
Our own file, annotated
# helloAISearch
> helloAISearch is an AI search visibility tool. It sends the questions
> your customers actually ask to ChatGPT, Claude, Gemini, and Perplexity ...
> Operated by PulseSpark.ai LLC, Pittsburgh, PA, USA.
Key facts:
- Category: AI visibility tool / AI search monitoring / AEO and GEO tracking
- Engines tracked: ChatGPT, Claude, Gemini, Perplexity ...
- Pricing: Lite $29/mo · Studio $59/mo · Pro $179/mo · Agency $449/mo ...
- Operator: PulseSpark.ai LLC, a Pennsylvania LLC (Pittsburgh, PA)
- Contact: hello@helloaisearch.com
## Pages
- [Pricing](https://helloaisearch.com/pricing): plans and what each includes
- [Compare](https://helloaisearch.com/compare): vs Otterly, Profound, ...Three choices worth copying. The category line lists every phrase someone might use for what we do, so an engine matching "tool to track brand mentions in ChatGPT" finds it. The pricing line states numbers, not "affordable plans". And the operator line names the legal entity and city, which is the fact engines most often get wrong about small companies.
One choice worth noting: the file is generated by a route handler, not a static file, so the URLs in it follow the environment. A staging site emits staging URLs; production emits production. A hand-edited file drifts.
What to leave out
- Marketing adjectives. "Best-in-class" tells an engine nothing it can repeat.
- Anything you would not want quoted. Engines quote llms.txt verbatim when they use it.
- Every page on the site. Link the ones that answer questions; the sitemap already lists the rest.
- Claims about competitors. Keep it to facts about you.
Does it help?
Honestly: it is early. Some crawlers fetch it, some ignore it, and no engine has said it affects citation. We keep ours because it is the canonical statement of our facts, because it is cheap, and because consistency of facts across sources is the one AEO signal every engine seems to reward. The file is one more source that agrees with the rest.