llms.txt

A markdown file at a domain's root that gives language models a curated index of the site's most important content.

term: llms.txt
category: protocols
short_def: A markdown file at a domain's root that gives language models a curated index of the site's most important content.
long_def: Proposed by Jeremy Howard (Answer.AI) on 3 September 2024. It mirrors robots.txt in spirit but is written for ingestion rather than exclusion: a concise, linkable map that helps agents find and prioritize content. A companion llms-full.txt inlines the full content.
see_also: markdown-twin agentic-web
etymology_origin: Proposed by Jeremy Howard, co-founder of Answer.AI, on 3 September 2024; specification maintained at llmstxt.org.
related_to: markdown-twin robots-txt agentic-web content-negotiation
contrast_with: Unlike robots.txt, which tells crawlers what NOT to fetch (exclusion), llms.txt tells models what to read first (curation for ingestion).
example: After Mintlify rolled out llms.txt support across its hosted docs in November 2024, thousands of docs sites — including Anthropic and Cursor — adopted the file.
source: https://www.answer.ai/posts/2024-09-03-llmstxt.html
status: proposed
why_it_matters: llms.txt is one of the lowest-effort, highest-signal agent-readiness steps: a single root file that curates what models ingest from your site.
sameAs: https://llmstxt.org/
bridge_entity: protocols/discovery/llms-txt
last_verified: 2026-06-15
md_twin: /glossary/llms-txt.md