Topic · GEO and AI visibility
llms.txt specification
The robots.txt of the AI era. Adoption is still under 5%.
DefinitionGEO and AI visibility
What is llms.txt specification?
llms.txt is a proposed plain-text manifest at /llms.txt that summarizes a website's content for large language model crawlers. The spec, hosted at llmstxt.org, defines a Markdown structure with sections of named links so models can ingest a curated map of the site.
Services that operate this topic
How Xpand Media runs it
Industries that care about this
Sector relevance
- VerticalB2B SaaSB2B SaaS in 2026 buys through ChatGPT and Perplexity before it ever sees Google. If your brand isn't cited in AI answers, demos don't book.
- VerticalAI-native companiesAI-native companies need GEO that names the model, the use case, and the developer adoption signal. Generic positioning loses to specific technical-buyer hooks.
FAQ
Common questions about llms.txt specification
Where does llms.txt live?
At the site root, at the path /llms.txt, served as text/plain. Following the spec, it uses Markdown structure with H1 (site name), blockquote summary, and named sections with link lists.
How is it different from robots.txt?
robots.txt controls crawler access. llms.txt curates content discovery — it tells the model what's worth reading, not whether it can read.
Do AI engines actually use llms.txt?
Adoption is partial. Anthropic publicly references it. OpenAI and Google have not confirmed dedicated parsers but treat it as crawlable plain text. Adding it is low cost and forward-compatible.
Should we also publish llms-full.txt?
Yes if the site has high-volume content. llms-full.txt is the long-form variant that concatenates page content, intended for offline LLM ingestion.
What format do links need to be in?
Standard Markdown: hyphen, space, bracketed link text, parenthesized URL, optional colon-prefixed description. Other formats (em-dash separated, plain URL lists) may fail spec parsers.
Talk to us