llms.txt Generator

Create a concise llms.txt draft that points AI agents to the most useful public resources.

Scope: drafting and reviewing the file itself. Whether crawlers can reach it belongs to robots-ai-crawler-audit; choosing which pages deserve to exist belongs to answer-engine-content-audit.

Workflow

1. Inspect the homepage, sitemap, docs, product/service pages, policy pages, and support pages.
2. Identify canonical high-value URLs and exclude thin, duplicate, private, or low-signal pages.
3. Group links by intent: overview, products/services, docs, support, policies, changelog, pricing, and contact.
4. Draft llms.txt with short descriptions for each section.
5. Note missing pages that would make the file more useful.
6. [EXPERIMENTAL] Check for emerging machine-readable discovery conventions (Markdown content-negotiation via Accept: text/markdown and Agential Resource Discovery manifests /auth.md or /.well-known/ard.json) — see references/checks.md. Clearly mark any findings as experimental draft standards.

Output