hydrateRoot vs createRoot
What you'll learn: The critical difference between mounting fresh React applications and rehydrating server-rendered HTML.
Two Ways to Boot React
When you start a React application in the browser, you have two entry points:
createRoot: Mounts a brand-new React tree with no assumptions about existing HTMLhydrateRoot: Attaches React to HTML that was already rendered on the server
The Kitchen Analogy
Think of createRoot like cooking a meal from scratch in your kitchen—you start with an empty pan and build everything yourself.
hydrateRoot is like receiving a pre-cooked meal from a restaurant and just reheating it—the food already exists, you're just making it hot and ready to serve.
When to Use Each
Use createRoot for purely client-side applications:
const root = createRoot(document.getElementById('root'));
root.render(<App />);
This completely replaces the DOM node's contents and builds the React tree from zero.
Use hydrateRoot after server-side rendering:
hydrateRoot(document.getElementById('root'), <App />);
This expects the HTML to already match what React would have rendered. React reads the existing markup and simply attaches event listeners and state management without rebuilding the DOM.
The Critical Distinction
hydrateRoot performs reconciliation between server HTML and the client React tree. If they mismatch, you'll see hydration warnings. createRoot doesn't care what was there before—it's a clean slate.
Key Takeaway: Use hydrateRoot when you've rendered HTML on the server with renderToString or renderToPipeableStream, and createRoot for client-only SPAs with no SSR.