How should ecommerce category pagination and infinite scroll expose products to crawlers?

Provide stable paginated URLs beneath the infinite-scroll experience.

Each result page should be directly requestable and expose standard HTML links to products. The user interface can append results dynamically, but search discovery should not depend on scrolling events.

Keep pagination canonical behavior consistent. Deeper pages generally should not all canonicalize to page 1 when they expose unique product sets needed for discovery. Avoid noindexing deeper pages automatically if doing so would remove the only crawl path to products.

Use real anchor links for next/previous or numbered states. Prevent sort-order and low-value filter parameters from multiplying each paginated state into many duplicates.

Ensure category page 1 links to important subcategories and products directly so priority inventory is not buried at excessive depth. XML sitemaps can support product discovery but should not replace internal links.

When products move between pages because of sorting or inventory changes, stable product URLs remain the canonical unit. Do not attempt to canonicalize product pages to category pages.

Test raw HTML, not just the browser experience. A crawler should be able to traverse from category page 1 into subsequent pages and then to product URLs without executing fragile UI actions.

The front end can feel infinite; the underlying information architecture should remain finite and linkable.