11 websites made with SvelteKit

Svelte is a compiler, not a runtime library. Components compile to direct DOM updates at build time, so no framework code ships alongside your application and there is no virtual DOM to diff. React does its work in the browser and ships the machinery to do it. In practice this means smaller bundles and less indirection between what you write and what runs — the trade is a much smaller ecosystem, which is the real decision for most teams.

Runes are the reactivity system introduced in Svelte 5, released in October 2024. $state, $derived, $effect and $props replace the previous model of reactive assignments and stores with fine-grained signals, closer to how SolidJS works. The benefit is that reactivity now behaves the same inside and outside components, which the old system did not. Migration is file by file, and an official script handles the mechanical parts.

Everything above the component layer: filesystem-based routing, load functions for data on server or client, form actions for mutations, layouts, error boundaries, and per-route rendering strategy — server-rendered, prerendered or client-only, chosen route by route rather than for the whole app. Adapters deploy the same codebase to Node, static hosting, Cloudflare, Vercel or Netlify. It is to Svelte roughly what Next.js is to React.

Around 45,000 live SvelteKit domains as of August 2026, a little over two percent of the web-framework market — eighth in its category. That is a real production user base and a small one next to React’s. Adoption has grown fastest in Europe. Developer satisfaction consistently runs ahead of usage: people who ship on it tend to want to use it again, which is why it keeps appearing near the top of framework surveys while sitting well down the usage tables.

When bundle size and runtime performance matter more than ecosystem depth, when the team is small enough that hiring for a specific framework is not a constraint, and when nobody needs a niche React library that has no Svelte equivalent. Next.js wins on ecosystem, hiring pool, hosting integrations and the sheer number of solved problems you can search for. SvelteKit wins on what actually ships to the browser and on how much of the framework you have to hold in your head.

Ecosystem size, mostly. Fewer component libraries, fewer third-party integrations, fewer people who have hit your problem before, and a smaller hiring pool. Svelte 5's runes also mean older tutorials and Stack Overflow answers describe a reactivity model that no longer applies, which makes learning from the existing body of material harder than the volume suggests. None of that is a technical limit on the framework — it is the cost of being the smaller option.