The canonical tag (rel="canonical") tells search engines which version of a page is the preferred one when duplicate or very similar URLs exist. This prevents Google from interpreting duplicate content as an attempt at manipulation and penalizing the site.
Common situations where a canonical is needed: pages accessible with and without www, with and without a trailing slash, with tracking parameters (UTMs), filtered versions in e-commerce and pagination pages. In all these cases, the canonical points to the "official" URL that should receive all the ranking credit.
Implementation is simple: add the tag to the page's head. In frameworks like Next.js, this can be done via the metadata API. The key is consistency: a page's canonical should point to itself (self-reference) or to the canonical version. Never leave pages without a canonical, especially on large sites.