#492 Serve llms.txt and llms-full.txt for AI engines and agents

[Anonymous]

Originally created by: athanclark

Closes [#482].

/llms.txt and /llms-full.txt answered 404.

/llms.txt (llmstxt.org), built by platform/lib/llms.js:

• Heading and summary: the site's name as the heading. The summary says the site runs OpenChronology and gives the one-sentence definition from [#481], then a paragraph on platforms, licence and exact times.
• Start here: Explore, the FAQ, the guide and Download.
• Who it is for: with the landing page on, each audience's page with its description.
• Help: every other help page, with its description from [#470].
• API: the reference (/docs/api) and the OpenAPI 3.1 description (/api/openapi.json).
• Public timelines: up to ten, featured first as guests browse them, none with sensitive-content labels, each with its description.
• Optional: /llms-full.txt and the source code.

/llms-full.txt: every help page's Markdown in one file. Each page sits under its title with its address, without the copyright line and description comment.

How they're served: both are generated from the help and the site's settings, so they follow the docs and each server's own name. They're served as text/plain, cached for an hour.

• Footer: every page's footer links llms.txt, since agents mostly find these files through links.
• Sitemap: pages.xml lists it.

Tests: next-platform checks both files' heading, summary, links and help entries, and the footer link. It passes here.

Docs: docs/sharing.md#search-engines.

🤖 Generated with Claude Code

https://claude.ai/code/session_011fFY8eruEi1xAkkMBni13B

[Anonymous]

Ticket changed by: athanclark

• status: open --> closed