The sitemap follows a published schema (<urlset> of <url> entries). Each entry has a <loc> and may have <lastmod>, <changefreq>, <priority>. A sitemap can also be an *index* (<sitemapindex>) that points to multiple sub-sitemaps — useful when you have more than 50 000 URLs or want to organise large content sets.
Sitemaps are discovered through robots.txt (Sitemap: line) or by submitting them in Search Console. They guide crawl prioritisation but do not grant a URL ranking. A URL declared in the sitemap but disallowed by robots is a contradiction Google will flag.
Modern frameworks generate the sitemap from the route tree. Next.js's App Router exposes app/sitemap.ts and generateSitemaps() for splitting; the URLs returned by those functions are the source of truth.