Docs pipeline

How this documentation set is rendered and deployed with the website.

1 min read

The docs pipeline has three layers:

  1. Markdown source in this repo, with the reading order kept in one navigation file beside it.
  2. The SecondPage website, which renders every page from that Markdown each time it deploys.
  3. Generated agent mirrors, served as static files beside the pages.

There is no separate build, sync, or publishing step. A Markdown change ships with the next website release at /docs/, and a page that is not in the navigation file fails the docs checks instead of going missing.

Commands

Regenerate the agent mirrors after changing a page title or summary:

Terminal
npm run agent:mirror --workspace @secondpage/product-docs

Check the public docs payload:

Terminal
npm run docs:check

This scans the Markdown and the generated agent files for placeholder text, stale tool names, source breadcrumbs, and internal implementation terms.

Search and Ask

The search index is built from the same Markdown on every deploy, one entry for each page and each section. Ask answers questions from these pages only and links the guide and section to read next. Its answers are written by AI and can be wrong, so every answer points back to the guide it came from. When Ask is unavailable, the closest guides from search are shown instead.

Still have a question?

Ask in your own words, or write to us at support@secondpage.cc.

Email us