← Semua picks

Stack Comparison Pick Astro

Astro vs Next.js vs Nuxt: which framework for a content site in 2026

Astro, Next.js, or Nuxt for your content site in 2026? A grounded comparison for small Indonesian dev teams on real budget constraints.

17 September 2026 · 8 minutes ·Use case: Choosing a framework for a content-heavy site — blog, docs, marketing — when team size is small and hosting budget is tight.
AstroNext.jsNuxt

Three good choices, one clear winner for content

Pick any content site project right now and you will run this comparison: Astro vs Next.js vs Nuxt. All three ship in 2026 with first-class content primitives. The question is which one costs you the least pain — in performance, in hosting bills, and in the hours your team spends fighting the framework instead of shipping.

This is that comparison, written for teams building blogs, documentation sites, landing pages, and marketing content — not SaaS dashboards. If you need auth, complex server state, and per-user data, this is not your decision guide.

Comparison at a glance

DimensionAstro 5Next.js 15Nuxt 4
Default JS sent to browserZero (opt-in islands)Medium (RSC helps, but easy to bloat)~15KB Vue runtime minimum
Content managementContent Layer (MDX, CMS, API)MDX via plugins, no native layerNuxt Content v3 (built-in, mature)
Deploy freeCloudflare Pages (unlimited BW)Vercel free tier (limits apply)Cloudflare Pages (Nuxt preset)
Cost at scaleNear-zero (static CDN)USD 20–60+/month on VercelNear-zero (static) or VPS
Team familiarity (Indonesia)GrowingLargest poolVue devs (smaller pool)
SEO / Core Web Vitals baselineExcellent out of the boxGood with tuningGood with tuning
Server-side app capabilityLimited (SSR mode exists, not primary)ExcellentExcellent
Learning curveLowMedium (App Router mental model)Low–medium

Astro 5

Why it wins for content

The zero-JS default is not a marketing line — it is a structural advantage. When Astro renders a page, the browser receives HTML. No JavaScript runtime, no hydration waterfall, nothing to parse before the first meaningful paint. For a blog reader on a 4G connection somewhere between the Tangerang exit toll and Serpong, that difference is measurable: 600ms vs 1.8 seconds on a content-heavy page is not academic.

Astro 5’s Content Layer is the feature that closed the gap with Nuxt Content. You define a content collection with a schema, point it at MDX files or a Sanity dataset or a REST endpoint, and Astro handles the type safety and build-time fetching. The config is explicit. There are no surprise revalidation cycles.

Cloudflare Pages + Astro is the most cost-efficient production setup for content sites in Indonesia right now. Static output, unlimited free bandwidth, CDN PoP in Singapore. A startup that bootstrapped out of a ruko office somewhere on Jalan Raya Serpong does not need a USD 40/month hosting bill for a company blog.

Where Astro falls short

Island components — dropping React or Vue components into an Astro page — work, but you are now maintaining two mental models simultaneously. If 40% of your pages need dynamic, interactive widgets, the island pattern becomes a coordination tax. That is the point where Next.js or Nuxt becomes the better base.

Astro SSR exists and runs on Cloudflare Workers or Node. It works. But the community answers for edge-case SSR problems are thinner than for Next.js or Nuxt. Do not pick Astro because SSR is possible — pick it because you want mostly-static output with occasional dynamic components.

Next.js 15

Strengths

Largest hiring pool in Indonesia, by a considerable margin. Every React developer applying for a mid-career role in 2026 has Next.js on their CV. If you are building a content site that will need handover — to a client’s in-house team, or to a junior you hire in six months — Next.js reduces risk on that front.

App Router with React Server Components is now mature enough to trust for new projects. RSC-only pages ship minimal JavaScript. When done right, a Next.js content page can match Astro’s performance output. The keyword is “when done right.”

Weaknesses

App Router caching has been revised four times since v13. In Next.js 15, the defaults changed again: fetch requests are no longer cached by default. That is the correct decision, but it means documentation and Stack Overflow answers from before mid-2024 are often wrong. A team unfamiliar with this history will hit confusing behavior on first deploy.

Vercel is the natural host for Next.js, and Vercel’s pricing scales quickly once you use ISR or Edge Functions at any real traffic volume. You can self-host on a VPS with Coolify — that solves the cost problem — but you lose some of the native Vercel integrations that justify choosing Next.js in the first place.

For a content site specifically, Next.js is heavier than the job requires. You are paying the App Router complexity cost without getting much back from it.

Nuxt 4

Strengths

If your team writes Vue, Nuxt 4 is the obvious choice over Astro — full stop. The Nuxt Content v3 module is genuinely excellent: queryable content collections, MDX support, a visual editor integration, and a document-driven routing mode that maps your file system to URLs cleanly. For a documentation site or a structured blog with multiple content types, Nuxt Content has more built-in tooling than Astro’s Content Layer.

Nuxt 4 on Cloudflare Pages with the Cloudflare preset gives you static output similar to Astro, with the full Vue ecosystem available. For a Vue-native team, the context-switching cost of learning Astro’s .astro syntax is a real consideration.

Weaknesses

The useAsyncData composable is a common stumbling block. In SSR mode, data fetched with useAsyncData is cached on the server by a key you provide. Get the key wrong — or forget to set it when calling from a component versus a page — and you get stale content on production that does not reproduce locally. This is a known issue that trips teams on first Nuxt SSR deploy.

Vue runtime ships to the browser by default, even for mostly-static pages. On Astro, a blog post page sends zero JavaScript. On Nuxt with static preset, it sends the Vue runtime (~15KB gzipped) plus page-specific hydration code. For LCP on slow connections, this is a real gap versus Astro.

Nuxt’s hiring pool in Indonesia is meaningfully smaller than Next.js. If you need to hire, expect fewer candidates with Nuxt experience.

Verdict

Choose Astro if:

  • Your site is primarily content: blog, docs, marketing, landing pages
  • Core Web Vitals and SEO are primary metrics, not secondary considerations
  • Your team can learn Astro’s syntax (it is approachable; give it two days)
  • Hosting budget is tight — Cloudflare Pages free tier is a real business decision
  • You do not need complex per-user server state

Choose Next.js if:

  • Team familiarity and hiring pool matter more than optimal performance
  • The content site will grow into a SaaS or need auth and dynamic routes later
  • You are self-hosting on a VPS and want the widest ecosystem of templates and libraries
  • Clients will maintain it themselves and recognize the Next.js name

Choose Nuxt if:

  • Your team already writes Vue and context-switching to Astro has a real cost
  • You want Nuxt Content’s queryable content collections and built-in tooling
  • You are comfortable with the Vue runtime overhead (~15KB) on static pages

For a pure content site in 2026, Astro is the pick. Not because it is new, but because its defaults are aligned with what a content site actually needs: fast HTML delivery, zero JavaScript overhead, and a hosting path that costs nothing until you have real scale. Next.js and Nuxt are excellent frameworks — they are just solving a broader problem than “ship fast, SEO-friendly content to Indonesian users on mixed connection speeds.”

Ditulis oleh Asti Larasati

// Pick Stack Comparison lain


← Semua picks RSS feed