On this page

08. What are preload, preconnect, prefetch, and prerender?

📄 Notesmedium#performance#resource-hints#browser#network

Short answer

They are resource hints that tell the browser about work it can start early. preconnect opens the connection (DNS, TCP, TLS) to another origin; preload fetches a resource the current page will need soon at high priority; prefetch fetches a resource for a likely next navigation at low priority; prerender loads and renders a whole next page in the background so navigation is instant (today done with the Speculation Rules API).

Explanation

Think of a dinner party. Preconnect is unlocking the front door before guests arrive. Preload is putting the main dish in the oven now because it is needed this evening. Prefetch is buying dessert for tomorrow if you have time. Prerender is setting tomorrow's whole table tonight.

HintScopePriorityTypical use
dns-prefetchNext connectionLowThird-party origins, cheap fallback for preconnect
preconnectCurrent pageHighFont CDN, API origin, image CDN
preloadCurrent pageHigh (by as type)LCP hero image, web fonts, critical CSS
modulepreloadCurrent pageHighJS modules and their dependencies
prefetchNext navigationLowest, idle timeNext route's JS chunk
prerender (Speculation Rules)Next navigationUses memory and CPUVery likely next page, like product to checkout

The old <link rel="prerender"> is deprecated; Chromium now uses <script type="speculationrules"> for prefetch and prerender.

Example

<head>
  <!-- Open connections early -->
  <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
  <link rel="dns-prefetch" href="https://analytics.example.com" />

  <!-- Needed for this page: fetch now -->
  <link rel="preload" href="/fonts/inter.woff2" as="font" type="font/woff2" crossorigin />
  <link rel="preload" href="/hero.avif" as="image" fetchpriority="high" />

  <!-- Probably needed on the next page -->
  <link rel="prefetch" href="/assets/checkout.4f2a.js" />

  <!-- Prerender likely next pages -->
  <script type="speculationrules">
    { "prerender": [{ "where": { "href_matches": "/products/*" }, "eagerness": "moderate" }] }
  </script>
</head>

React 19 exposes the same hints as functions, and deduplicates them into the document head:

import { preconnect, prefetchDNS, preload, preinit } from "react-dom";

function ProductPage() {
  preconnect("https://cdn.example.com");
  prefetchDNS("https://analytics.example.com");
  preload("/fonts/inter.woff2", { as: "font", type: "font/woff2", crossOrigin: "anonymous" });
  preinit("/widgets/reviews.js", { as: "script" }); // fetch and execute
  return <h1>Product</h1>;
}

Frameworks also prefetch for you: Next.js <Link> prefetches routes in the viewport, and React Router's <Link prefetch="intent"> (framework mode) prefetches on hover.

Common mistakes

  • Preloading too many files; everything becomes high priority and competes with the real critical resources.
  • Preloading fonts without crossorigin, which causes a second download.
  • Preconnecting to many origins; each connection costs CPU and is closed if unused within about 10 seconds.
  • Prerendering pages with side effects (analytics, "mark as read") without handling the prerendering state.

Follow-up questions

  • How do you find what to preload? Look at the LCP element and late-discovered resources in Lighthouse or the Performance panel's request waterfall.
  • What is fetchpriority? An attribute (high/low/auto) that adjusts one resource's priority, e.g. the hero image.
  • Prefetch vs preload in one line? Preload is for this page now; prefetch is for a future page later.

Quick recap

  • preconnect: warm up the connection to an origin.
  • preload: fetch a current-page resource early at high priority.
  • prefetch: fetch next-page resources when idle.
  • prerender: render the next page in advance (Speculation Rules).
  • React 19 offers preconnect, prefetchDNS, preload, preinit.