llms.txt is a plain-text file at your domain root that tells AI agents and crawlers where to find your most important content. Think of it as a curated reading list for machines.
Why llms.txt matters for GEO
Answer engines and IDE agents do not browse like humans. They need:
• A clear starting point (homepage, docs index, API reference)
• Short descriptions of what each link contains
• Hints about which pages are authoritative vs supplementary
Without llms.txt, agents guess from your sitemap or scrape navigation — often missing the page that answers the user's question.
Format
llms.txt uses Markdown-style headings and link lists:
# Your Company
> One-line description of what you do.Docs
• Getting started: onboarding guide
• API reference: REST endpoints
Product
• Pricing: plans and limits
Keep it concise. Agents read the whole file in one fetch.What to include
1. Site summary — One blockquote describing your product or publication.
2. Key sections — Docs, product, help, legal — with 3–8 links each.
3. Descriptions — After each link, a short clause on what the page contains.
4. Optional: contact — Support email or status page for agents that need escalation.
Skip marketing fluff. Machines want signal.
ADN publishes its own llms.txt
We eat our own dogfood. ADN's marketing site serves llms.txt at https://www.agent-delivery.network/llms.txt — including segment pages, help guides, and this learn section.
How ADN uses llms.txt
When ADN classifies a request as developer_agent or answer_engine, llms.txt is one of the discovery signals. Sites with a well-formed llms.txt file:
• Rank higher in the free efficiency scan's GEO signals
• Give IDE agents a faster path to authoritative docs
• Reduce wasted token spend on navigation pages
Check yours today
Run the free scan at https://www.agent-delivery.network/#scan. It reports whether llms.txt is present and reachable. If it is missing, add one before worrying about Markdown conversion or MCP tools.