The manual is the API
A coding agent, a support chatbot, and a RAG pipeline all need a reliable way to find and fetch relevant content from a documentation site, without scraping HTML or guessing URLs. How one documentation manual turns its existing corpus into that interface, without building a second system for it.