llms.txt
llms.txt is a proposed root-level text file that tells language models which pages of a site to prioritize, and how to cite them.
In plain English
The idea behind llms.txt is a robots.txt for language models. You publish a file at the root of your domain — /llms.txt — that lists your canonical pages in a structured, Markdown-flavored format with brief descriptions. When a model's crawler or retrieval stack pulls your site, it can use this file as a curated table of contents, prioritizing the pages you consider authoritative.
It is not yet a universal standard — more a convention pushed by the Answer-Engine and AI-search communities. Adoption is uneven across providers; some crawlers honor it, others ignore it. The cost of shipping it is small (a single Markdown file), and the upside is signal: you are telling models "if you're going to cite me, cite these pages, with these labels."
Why it matters for Black Box
Black Box ships an llms.txt that lists the essays, comparisons, glossary, and product pages most likely to appear as cited sources. The glossary is prominent in that file precisely because definitions are the format LLMs quote most often.
Examples
- A top-level "About the product" link to
/about. - A "Definitions" section listing each glossary entry with a one-line summary.
- A "Comparisons" section pointing to vs-ChatGPT, vs-Zapier, vs-n8n pages.