Next.js is the default answer for React teams. Astro is the framework that keeps winning benchmarks for pages made of text, images and links. If what you are building is a content site (a blog, a marketing site, documentation, a template you will hand to clients), the choice comes down to one question: does this page need a JavaScript runtime to exist?
Short version: Astro for content, Next.js for applications. The rest of this post is the reasoning, plus a way to check it on your own sites.
What Ships to the Browser
A Next.js page is a React page. Even when the route renders nothing interactive, the visitor downloads React, ReactDOM and the Next.js client runtime before your own code starts. Server Components in the App Router moved a lot of work to the server and cut the per-page bundle, but the shared client runtime still ships on every route. Run next build on a fresh project and read the “First Load JS” column: that is your floor before a single component of yours.
Astro’s floor is zero. Pages are rendered to HTML at build time and no client framework is included unless you add an island with a client:* directive. A page of prose ships HTML and CSS and nothing else, which is why “100 Performance” is the normal result rather than an achievement.
For a twenty-page marketing site this compounds: twenty routes each paying the runtime tax on first load, versus twenty routes that are just documents.
How to Measure It Instead of Trusting a Table
Framework comparisons love a benchmark table, and every one of them was produced on someone else’s site. Do your own in a minute: take any Astro site and any Next.js site you know, and run them side by side through the free website comparison tool. It runs Google’s PageSpeed test on a simulated phone for each and shows the four Lighthouse scores and the on-page checks next to each other. Total Blocking Time is the number to watch; it is where a client runtime shows up first.
SEO: Both Work, One Has Fewer Moving Parts
Google can crawl both. The practical differences:
Next.js renders on the server, so crawlers see full HTML. Hydration then runs on the client, and anything that changes layout during hydration counts against Core Web Vitals. The Metadata API covers titles and descriptions; structured data, Open Graph images and sitemaps are yours to assemble.
Astro emits static HTML, so view-source is exactly what the crawler sees. There is no hydration step on a page without islands, so there is nothing to shift. Because content lives in Content Collections with a schema, a layout can generate JSON-LD, canonical tags and sitemap entries from the same validated data that renders the page.
Both support canonical URLs, sitemaps and robots directives. Astro’s edge is predictability: a document that does not change after load is the easiest possible thing for a search engine to evaluate.
Developer Experience
Where Next.js is stronger:
- The React component ecosystem, which is enormous
- Route handlers and middleware for backend logic in the same project
- Mature patterns for auth, sessions and per-request rendering
- First-class hosting on Vercel
Where Astro is stronger:
- A simpler mental model: components are HTML with a frontmatter script
- Islands can be React, SolidJS, Vue, Svelte or Preact, on the same page if you want
- Content Collections with Zod validation, built in
- Static output that deploys to any host that serves files
If your team lives in React and the product needs API routes, auth and heavy client state, use Next.js. If the product is pages, Astro removes a whole layer you would otherwise be maintaining.
Lock-In
Choosing Next.js is choosing React for every page, every component, every interactive element, and React’s direction becomes your direction. The Server Components transition showed what that costs when the model shifts underneath a codebase.
Astro is framework-agnostic. Amrify Pro ships its interactive widgets (estimate calculator, quote wizard, reviews carousel) as SolidJS islands because SolidJS compiles to direct DOM updates with no virtual DOM runtime, so a widget carries its own logic and little else. The page templates around those widgets do not know or care which framework the island uses. Swapping it would not touch a layout.
Content
Build Behavior
Astro builds are mostly file emission: render each route, write HTML, copy assets. There is no client bundle to produce for a page without islands, so adding pages adds roughly linear, cheap work. A Next.js build has to compile and bundle React components for the client on top of rendering, and that is the part that grows with the size of your component tree.
How much that matters depends on how often you deploy. A monorepo of many sites that deploys several times a day feels the difference; a blog that deploys weekly does not.
Choose Next.js When
- You need server logic in the same project: route handlers, auth, database access
- The product is an application with real client-side state
- The team is React-native and the ecosystem is the draw
- You want middleware running on every request
Choose Astro When
- The site is content: blogs, marketing, documentation, portfolios, local business sites
- Performance and Core Web Vitals are requirements, not aspirations
- You are building templates or a multi-site system for clients
- You want interactivity as opt-in islands rather than a framework on every page
Where Amrify Landed
Amrify is built on Astro because the job (business sites and agency templates) is content all the way down. The free Starter takes the idea to its end: no JavaScript framework at all, with the mobile menu, FAQ accordion and dark mode handled by the platform, and Lighthouse 100 in all four categories enforced in CI. Pro adds SolidJS islands only where a widget earns them.
If you are weighing the two, build one page in each. The difference in what you are responsible for shows up before lunch. Coming from WordPress rather than React? The Astro vs WordPress comparison is the better starting point.
Settle it with numbers, not opinions. Put an Astro site and a Next.js site through the free You vs. The Competition tool and compare the four Lighthouse scores side by side; no account, about thirty seconds. Then, if the Astro column wins, the free Starter is where that column comes from.
