Prefetching Data#

prefetch runs a loader and fills its cache ahead of navigation, so the destination route renders immediately when the user arrives. Use it when you can predict the user's next move (hover on a link, scroll into view, idle after page load) and the loader is non-trivial.

Basic usage#

import { prefetch } from 'hono-preact';
import { serverLoaders } from './pages/movie.server.js';

const movieLoader = serverLoaders.default;

// On link hover, kick off a fetch for /movies/42 so the route is ready when clicked.
function onMovieLinkHover(id: number) {
  prefetch(movieLoader, { url: `/movies/${id}`, route: '/movies/:id' });
}

The result is written into the loader's cache (when one is configured) and the next navigation reads from cache instead of refetching.

Common patterns#

Hover prefetch is the typical case: bind to onMouseEnter on a link and call prefetch with the destination URL.

Idle prefetch is useful for the next-most-likely route. Call prefetch from a requestIdleCallback after the current page settles.

Programmatic prefetch works anywhere: a router transition listener, a search-as-you-type debounce, an intersection observer for cards scrolling into view.

prefetch checks the loader's cache first and returns immediately when the destination is already warm. A hover handler firing repeatedly (mouse over → off → back over the same link), an intersection observer re-entering visibility, or duplicate idle-callback scheduling will not issue redundant network requests. The cache key combines path with whatever searchParams the loader's params option declares, so two prefetches for /movies/41 and /movies/42 each hit the network exactly once.

Why pass url and route?#

Loaders receive a RouteHook location ({ path, searchParams, pathParams }). When your loader reads location.pathParams.id or location.searchParams.q, prefetching with no location would crash or return the wrong data. Pass the URL the user is about to navigate to, plus the route pattern, and prefetch derives a complete RouteHook for you:

• url populates path (trailing slash stripped, root preserved) and searchParams.
• route is the matching route pattern (e.g. /movies/:id); when present, pathParams is derived by matching url against it.

If you don't read any of those fields, omit both arguments and prefetch(loader) still works.

Options#

See also#

• Server Loaders: the loader cache prefetch fills.
• Link Prefetch: the declarative data-prefetch attribute.
• Loading States: what prefetching lets you skip.