Static Site Generators 2026 Head-to-Head — Hugo · Eleventy · Astro · MkDocs · Docusaurus · Mintlify · Starlight · Nextra · VitePress · Jekyll · Zola Deep Dive

Prologue — SSGs Didn't Die, They Got Redefined
1. The Comparison Axes — What to Actually Look At
2. Hugo — Go's Speed, 11 Years of Stability
3. Eleventy (11ty) — The JavaScript Camp's Minimalist
4. Astro — The New Content-First Standard
5. MkDocs (Material) — Python Camp's De Facto Standard for Docs
6. Docusaurus — Meta's React + MDX Docs Framework
7. Mintlify — "We Handle Design" for SaaS Docs
8. Starlight — Astro's Rising Docs Heavyweight
9. Nextra — The Next.js Docs Theme
10. VitePress — Vue Camp's MkDocs
11. Gatsby, Jekyll, Zola — The Surviving Pack
12. Category Matrix — At a Glance
13. Scenario Matrix — Which Tool for Which Job
14. The AI-Docs Era — New Variables
15. Migration Cost — What Switching Really Means
16. Epilogue — Decision Checklist and Anti-Patterns
References

Prologue — SSGs Didn't Die, They Got Redefined

Spring 2026: someone tweeted "SSGs are dead, everyone moved to Next.js Server Components." The tweet reached 10,000 people and 9,000 disagreed. The real landscape is the opposite. Static site generators — SSGs — haven't died over the last five years; they have split into two branches and gotten stronger.

One branch is content sites: company blogs, marketing sites, personal portfolios, newsletter archives. Hugo builds 1,000 pages in seconds with Go's speed. Eleventy 3 has redefined what minimalism in the JS camp looks like. Astro created a category called "content-first." For this branch the priorities are SEO, page speed, RSS, i18n, and a sane Markdown workflow.

The other branch is documentation sites: OSS library docs, SaaS API references, internal wikis, changelogs. MkDocs Material became the de facto standard in the Python camp. Docusaurus, Meta's React + MDX combo, powers tens of thousands of OSS projects. Mintlify said "pay us and don't worry about design" and won the SaaS API doc market. Astro's Starlight has emerged as a new contender. Nextra is the Next.js doc theme, and VitePress is the Vue answer.

And then there are the dying or dead camps. Gatsby — once the React SSG, effectively in maintenance mode after the Netlify acquisition. Almost nobody chooses Gatsby for a new project in 2026. Jekyll — Ruby's original SSG and the GitHub Pages default. Still running, but momentum is gone. Hexo — once big in the Node camp, now mostly hanging on in Korea and China. Zola — a fast Rust SSG with momentum, but the ecosystem is thin.

This post compares eleven SSGs along the same axes. After the comparison we look at four scenarios — company blog, personal blog / portfolio, OSS library docs, SaaS API docs — and ask which tool to pick. Then we cover the new 2026 variables: AI search, translation, and what llms.txt changes about doc-site choice.

Tools in the same category can produce very different outcomes once content models and operational models diverge. Switching SSGs after one year of operation costs roughly ten times what initial adoption cost.

Prices and feature numbers change fast. Every number here is from May 2026 and the focus is structural differences. Six months later the numbers will move, but the decision frame should still hold.

1. The Comparison Axes — What to Actually Look At

We decompose the eleven tools along nine axes. The axes themselves are the decision frame.

Axis 1 · Category fit. Is the tool strong for content sites, doc sites, or both? Hugo and Eleventy are content-strong; MkDocs, Docusaurus, and Mintlify are docs-strong; Astro covers both; Next.js itself is full-stack. Ignore category at your peril.

Axis 2 · Build speed. The build-time curve as the page count grows. Hugo is overwhelmingly fast (Go compile plus concurrency). Zola comes next (Rust). All the JavaScript tools compete in roughly the same band. Past 1,000 pages the gap widens sharply.

Axis 3 · Content model. Does it accept only Markdown, MDX (JSX in Markdown), AsciiDoc, or reStructuredText? This decides freedom of expression. MDX lets you embed React components inside prose, which is great for interactive widgets but couples content to code and makes migration hard.

Axis 4 · Theme and design freedom. Is it free-form or theme-led? MkDocs Material, Mintlify, and Starlight ship a near-fixed design. Hugo, Eleventy, and Astro start from a blank canvas. Freedom and time-to-launch are a trade-off.

Axis 5 · Search. Built-in, external, or AI-powered? MkDocs Material ships lunr.js client search. Docusaurus recommends Algolia DocSearch. Mintlify provides AI search out of the box. AI-search SaaS like Inkeep and Kapa.ai are quickly becoming standard add-ons.

Axis 6 · Internationalization. Multi-language routing and translation workflow. Docusaurus, Astro, and Hugo are first-class. Eleventy needs a plugin. MkDocs needs the i18n plugin. Jekyll is hard. If you need multilingual, the candidate list shrinks by half.

Axis 7 · Deployment / hosting model. Self-host static output anywhere, or pay for SaaS? All the OSS tools deploy to GitHub Pages, Cloudflare Pages, Vercel, and Netlify. Mintlify is SaaS-only — no self-host. Self-host versus SaaS is a cost / control / speed trade.

Axis 8 · Pricing. OSS, partly paid, or full SaaS. Mintlify is paid: Starter free, Pro around 150 dollars per month, Growth around 550 dollars per month (May 2026, subject to change). MkDocs Material is OSS with an Insiders sponsor tier. The rest are OSS plus your own hosting.

Axis 9 · AI integration / future fitness. Is the content model LLM-friendly? Is it easy to attach AI search? Does it support emerging standards like llms.txt? Mintlify leads, with Docusaurus and Starlight catching up fast. In 2026, one of the largest traffic sources for doc sites is no longer search engines — it is LLMs.

Keep these nine axes in mind. Now we look at the tools one by one, each in the same shape.

2. Hugo — Go's Speed, 11 Years of Stability

Category and strengths. A Go-built SSG. First release in 2013, currently in the v0.140+ series in 2026. One word summary: speed. It builds 10,000 pages in under a minute. That is a band the JS camp simply cannot reach. It also ships as a single binary with no external dependencies, so CI caching and reproducibility are trivial.

Content model. Markdown plus YAML/TOML frontmatter. Directory-based routing — drop a Markdown file under content/posts/ and it appears in the site tree. Multilingual handled via separate content.ko/ and content.en/ directories or filename suffixes. The shortcode system lets you embed widgets in prose, but it is less free than MDX.

Themes. Rich official theme gallery: PaperMod, Hello Friend, Stack, Doks. Writing your own theme is also straightforward — the Go template syntax is unfamiliar but consistent once learned.

Search. No built-in search. Bolt on Fuse.js, Lunr.js, Pagefind, or Algolia. Pagefind is the most popular pick in 2026 — it builds a static index at build time and searches it client-side.

Weaknesses. Go template syntax (the .Title style) feels alien to JavaScript developers. No JSX or MDX, so embedding React components in prose is impossible. The build-speed advantage is dramatic past 10,000 pages, but between 100 and 500 pages the perceived speed gap with Astro is small.

Typical users. Large company blogs, big news sites, newsletter archives. It is common to see a single team running hundreds of thousands of Hugo pages. 1Password Docs, parts of Smashing Magazine, and a vast number of personal blogs all run on Hugo.

3. Eleventy (11ty) — The JavaScript Camp's Minimalist

Category and strengths. A JS SSG that landed in 2018. Eleventy 3.0 shipped in 2024 with ESM, edge functions, and faster builds. The core philosophy is "no framework" — your site is your site, and Eleventy is merely a Markdown-to-HTML converter. JAMstack purists love it.

Content model. Markdown, Liquid, Nunjucks, Handlebars, EJS, JavaScript, TypeScript, HTML — Eleventy accepts almost any template language. You can even mix template engines page by page within one site. That freedom is the brand.

Themes. Almost no official themes. A handful of starter templates and most users roll their own. The eleventy-base-blog and eleventy-excellent starters are popular.

Search. None built in. Bolt on Pagefind, Lunr.js, or Algolia. The official Eleventy docs themselves use Pagefind.

Weaknesses. No React/JSX/MDX-style interactive component workflow. You write your own Web Components or hand-roll an islands pattern. So Eleventy is ideal for "just Markdown, please" sites but limited when prose needs to include chart widgets.

Typical users. Personal blogs, small company sites, OSS project landing pages. Parts of Google Chrome Developer, parts of web.dev, and many independent JS-developer blogs run on 11ty.

4. Astro — The New Content-First Standard

Astro is covered in depth in a separate post; here is the comparison-friendly summary.

Category and strengths. Launched in 2021. Astro 5.0 shipped in 2024, and the v6.x line followed in late 2025. Content-first, islands architecture (partial hydration), first-class MDX, and multi-framework (React, Vue, Svelte, Solid coexisting in one site) are the core ideas. By default the build ships zero JavaScript, with interactive components hydrated only where needed.

Typical users. One of the most popular SSGs in 2026. Company blogs, OSS landings, marketing sites, and now — via Starlight (chapter 8) — doc sites too.

Weaknesses and a closer look are in the dedicated post.

5. MkDocs (Material) — Python Camp's De Facto Standard for Docs

Category and strengths. A Python-built static site generator. MkDocs itself is a very simple tool, but the Material for MkDocs theme effectively redefined the category: "documentation site equals Material for MkDocs" became conventional wisdom. Clean design, solid search, rich extensions — and above all easy to write. A single YAML file holds the site structure and configuration.

Content model. Markdown plus frontmatter. Directory structure equals site structure. List your nav in mkdocs.yml or use the awesome-pages plugin to auto-arrange.

Themes. The Material theme is the de facto default. An Insiders sponsor tier exists with extra features, but the core features eventually land in the free version (time-delayed release).

Search. Client-side lunr.js search out of the box. Multilingual search too. Search quality is above average and works without configuration.

Weaknesses. No React/JSX/MDX. Interactive widgets in prose require vanilla JS or iframes. The content model is locked into Markdown plus the pymdown-extensions ecosystem.

Typical users. 60 to 70 percent of Python OSS library docs run MkDocs Material: FastAPI, Pydantic, Pelican, and Material's own docs. Infrastructure OSS like OpenStack, Cilium, and k3s also use it.

6. Docusaurus — Meta's React + MDX Docs Framework

Category and strengths. A React-based docs SSG built by Meta (then Facebook) in 2017. v2 in 2022, the v3 line stabilized in 2024 and 2025. React + MDX + i18n + versioning are all first-class. It is the most-tested option for an OSS library that needs to ship Korean, Japanese, Chinese, and English docs while keeping v1, v2, and v3 docs live.

Content model. First-class MDX. Embed React components in prose, so an interactive demo can sit next to a code example. Sidebars are declared in sidebars.js as JavaScript, so dynamic sidebar generation is possible.

Themes. The official Classic theme is the default. Design is plain but solid. Customization uses the swizzling model — copy a theme component into your project and modify it.

Search. Algolia DocSearch is the official recommendation, free for OSS. Other integrations include lunr, typesense, and Mendable (AI). In 2026, AI-search integrations like Kapa.ai and Inkeep are growing fast.

Weaknesses. Builds slow down past 1,000 pages, scaling into minutes — a common React-SSG weak spot. Turning on i18n, versioning, and search together makes configuration complex quickly.

Typical users. React Native, Jest, Babel, Redux, Yarn, Storybook, Prisma — the standard for Meta-adjacent and major OSS docs. SaaS API docs commonly use it too (Supabase was on Docusaurus before switching to a custom build).

7. Mintlify — "We Handle Design" for SaaS Docs

Category and strengths. A paid managed documentation platform that launched in 2022. Their insight: the cost of self-hosting an OSS SSG often equals a designer's salary. Design is polished from minute one, AI search is built in, and pointing an OpenAPI spec at it auto-renders an API reference. By 2026, Mintlify has become one of the de facto standards for SaaS API docs.

Content model. MDX. A rich proprietary component library: Accordion, Card, Tabs, Steps, and more. One mint.json (or docs.json) file holds site structure, theme, and SEO meta.

Themes. Mostly fixed. You change colors, logo, and fonts. That sounds limiting but is the point — no designer needed, looks polished immediately.

Search and AI. Search plus an AI chatbot that synthesizes answers from search results is built in. llms.txt is auto-generated. OpenAPI auto-renders. An API playground is bundled. It is the most AI-forward tool in the category.

Weaknesses and pricing. SaaS only — no self-host. Pricing rises quickly. As of 2026, Starter is free but caps pages and members; Pro starts around 150 dollars per month, Growth around 550 dollars per month (exact numbers move). Larger orgs with large doc sites can easily exceed 10,000 dollars per year.

Typical users. Anthropic Docs, Mintlify's own docs, Cursor, Linear API, and many YC startups' API docs. The bundle is for companies that want "ship fast, look great, AI search included" without building it themselves.

8. Starlight — Astro's Rising Docs Heavyweight

Category and strengths. The official documentation theme by the Astro team, released in 2023. It is a full-stack SSG built on Astro, but from the user's perspective it is "one command to start a docs site." It matured rapidly through 2024 and 2025, and is now a serious alternative to Docusaurus and MkDocs Material in 2026.

Content model. MDX plus Astro components. Type-safe frontmatter via Astro's content collections. Sidebars, search, i18n, and versioning are all built in.

Themes. Design is clean from the start. Light/dark mode, sidebars, and search UI are pre-configured, so a real "five-minute start" is genuinely possible.

Search. Pagefind is built in (static index plus client search). Algolia, Inkeep, and Kapa.ai integrations are easy. The official Astro docs include an AI search integration guide.

Weaknesses. The ecosystem is still thinner than Docusaurus. Fewer plugins, fewer large OSS projects using it as a base. That gap is shrinking each quarter.

Typical users. Parts of Cloudflare Wrangler docs, Bun (partially), and dozens of rising OSS projects. Starlight is increasingly the default for new OSS projects shipping in 2026.

9. Nextra — The Next.js Docs Theme

Category and strengths. A Next.js-based docs theme from Vercel. A natural fit for orgs already running Next.js sites that want docs in the same stack. v3 in 2024, v4 in 2025. A hybrid SSG/SSR running on Server Components.

Content model. MDX plus Next.js App Router. Sidebars are defined per-directory through _meta.json files. Search via FlexSearch or Algolia.

Themes. Docs theme and Blog theme. Design is plain but the upside is full access to every Next.js capability.

Weaknesses. The Next.js dependency is heavy. Pulling the entire Next.js build chain in for plain docs is often overkill. Build time is also slower than lighter SSGs.

Typical users. SWR, Turborepo, and other Vercel-adjacent OSS. Internal docs at companies already running Next.js.

10. VitePress — Vue Camp's MkDocs

Category and strengths. A Vite-based docs SSG by the Vue core team. Feels like MkDocs Material rebuilt for the JavaScript camp: a clean default theme, fast builds, and Markdown extensions that let you embed Vue components in prose. v1 in 2024, v2 in 2025.

Content model. Markdown plus Vue components (not MDX, but similar capability). Sidebars from config. Builds run on Vite — fast.

Themes. Vue's official-docs style is the default. The design is solid and consistent. Customization is via Vue component overrides.

Weaknesses. Little momentum outside the Vue camp. You cannot embed React components in prose. i18n landed in late 1.x but is less mature than Docusaurus.

Typical users. Vue.js, Vite, Pinia, parts of Nuxt, Vitest. The standard for Vue OSS. Non-Vue orgs rarely pick it.

11. Gatsby, Jekyll, Zola — The Surviving Pack

Gatsby — effectively over. The starting point of React SSGs in 2015. A GraphQL data layer, a plugin ecosystem, and great marketing. It was the React SSG standard from 2020 to 2022. Then Next.js absorbed SSG/SSR, Astro took over content sites, and Netlify acquired Gatsby and put the company into effective maintenance. New adoption dropped sharply after 2024. Existing Gatsby sites continue; new projects rarely pick it in 2026.

Jekyll — alive but stagnant. Built in 2008 by GitHub co-founder Tom Preston-Werner. The default SSG for GitHub Pages, where countless personal blogs and docs live. Stable but momentum is gone. Ruby dependency is a friction point, there is no MDX/JSX/Vue, and builds are slow. Existing Jekyll sites are fine to keep; few recommend it for new work.

Zola — Rust's rising star. A Rust-built SSG: single binary, very fast builds, Tera templates. Sits closest to Hugo on the map. The downside is a thin ecosystem — far fewer themes, plugins, and community members than Hugo. Momentum is real and the audience is "Hugo feels too big, give me something smaller."

Hexo — surviving in Korea and China. Once a Node SSG heavyweight. Still in use in a handful of regions but no global momentum.

12. Category Matrix — At a Glance

Tool	Category	Language	Content model	Build speed	Design freedom	Search	i18n	Pricing
Hugo	Content	Go	Markdown + Shortcode	Very fast	High	External	First-class	OSS
Eleventy 3	Content	JS	Markdown + multi-template	Fast	Very high	External	Plugin	OSS
Astro	Both	JS/TS	MDX + Astro	Fast	High	External / Pagefind	First-class	OSS
MkDocs Material	Docs	Python	Markdown	Moderate	Low (fixed theme)	Built-in	Plugin	OSS + Insiders
Docusaurus	Docs	JS/TS	MDX	Slow-ish	Moderate	Algolia	First-class	OSS
Mintlify	Docs	(Managed)	MDX	Instant	Low (fixed theme)	AI built-in	Supported	Paid SaaS
Starlight	Docs	JS/TS	MDX + Astro	Fast	Moderate	Pagefind built-in	First-class	OSS
Nextra	Docs	JS/TS	MDX	Moderate	Moderate	FlexSearch / Algolia	First-class	OSS
VitePress	Docs	JS/TS	Markdown + Vue	Fast	Moderate	Built-in	Partial	OSS
Gatsby	Both	JS	MDX (separate)	Slow	High	External	First-class	OSS (declining)
Jekyll	Both	Ruby	Markdown + Liquid	Slow	High	External	Hard	OSS
Zola	Content	Rust	Markdown + Tera	Very fast	High	External	First-class	OSS

This is the big picture. Each tool has scenarios where it wins, so read it alongside the next chapter's scenario matrix.

13. Scenario Matrix — Which Tool for Which Job

Define four scenarios and pick first and second choices for each.

Scenario A · Personal blog or portfolio. Requirements: Markdown authoring, fast builds, free deploy to GitHub Pages or Cloudflare Pages, RSS, well-tuned SEO, dark mode.

First pick: Astro — rich starter templates, clean content collections, lightweight output.
Second: Hugo — once you have many posts, the build-speed gap shows.
Third: Eleventy — for JS-camp minimalists.

Scenario B · Company blog or marketing site. Requirements: design freedom, interactive widgets, multilingual, CMS integration (Notion / Sanity / Contentful), Vercel or Cloudflare deploy.

First pick: Astro — MDX, multi-framework, i18n, and CMS integrations are all strong.
Second: Next.js plus a content solution — when you need full-stack.
Third: Hugo — high design freedom and proven i18n.

Scenario C · OSS library docs. Requirements: easy Markdown authoring, i18n, versioning (v1, v2, v3 in parallel), search, OSS-friendly licensing, easy community contributions.

Python camp: MkDocs Material — the de facto standard. Picking otherwise usually causes regret.
React/JS camp: Docusaurus — proven by Meta and the OSS community.
New projects: Starlight — better when you want small and fast.
Vue camp: VitePress — no other real choice.

Scenario D · SaaS API docs. Requirements: OpenAPI auto-render, API playground, AI search, fast and polished, save designer time.

First pick: Mintlify — trading dollars for time. Design, AI search, and OpenAPI automation all bundled.
Second: Docusaurus plus redocusaurus plus Algolia plus Kapa.ai — self-hosted with integration. Cheaper but high-touch.
Third: Starlight with an OpenAPI plugin — momentum is building here.

Scenario E · Internal wiki or changelog.

First pick: MkDocs Material or Docusaurus — both proven. Pick by team language: Python team → MkDocs, JS team → Docusaurus.
Second: Starlight — for a small team starting fast.

14. The AI-Docs Era — New Variables

Building a docs site in 2026 means more than "make pages for humans to read." LLMs read them and answer. Three new variables.

Variable 1 · llms.txt and llms-full.txt. Putting an llms.txt at the site root lets LLMs understand the structure quickly. It became a de facto standard in late 2025. Mintlify auto-generates it. Docusaurus and Starlight have plugin support. Without it, ChatGPT and Perplexity will index the site poorly.

Variable 2 · On-site AI search. Page search alone is no longer enough. When a user asks "how do I do X?" an LLM should synthesize an answer from the docs. Mintlify includes this by default. Otherwise you bolt on Kapa.ai, Inkeep, Mendable, or Algolia AI. Cost runs roughly 200 to 1,000 dollars per month. Worth considering once monthly doc traffic crosses 100,000.

Variable 3 · Structured content. LLMs cite pages well when title, h1, code block languages, meta descriptions, and Open Graph tags are accurate. Whether the SSG auto-emits this metadata or makes the author hand-write it decides the site's LLM-friendliness. Mintlify, Starlight, and Docusaurus are strong here; bespoke builds are often weak.

In 2026, 30 to 50 percent of doc-site traffic arrives via LLM-synthesized answers. Direct human pageviews drop while LLM-citation pageviews rise. Sites that ignore this shift will be buried within six months.

15. Migration Cost — What Switching Really Means

Switching an SSG you have operated for a year costs about ten times what initial adoption did. That is why the first choice matters. Break the migration work down:

Content conversion. Markdown-to-Markdown is comparatively easy, but mismatched frontmatter field names, shortcodes, and MDX components require a one-off conversion script. Several days for large sites.
URL preservation. If old URLs change, SEO collapses. Translate routing rules carefully and keep a year of redirects.
Search index rebuild. Rebuild Algolia or your homegrown index. Search quality changes and users complain.
Design fidelity. Pixel-perfect reproduction is impossible. Some pages drift slightly, and for a while users say "the old version looked better."
Build pipeline. CI config, caching, deploy hooks, and preview environments all need to be rebuilt for the new tool.
Author retraining. Existing authors must learn the new Markdown extensions and MDX components. Budget a week.

The best migration is none. If you must move, stay within the category (MkDocs Material to Starlight is comparatively easy, MkDocs to Docusaurus is hard, MkDocs to Mintlify has some auto-migration tooling but always needs review).

16. Epilogue — Decision Checklist and Anti-Patterns

Checklist before you pick

What category is the site? Blog or marketing site, technical docs, or both? The answer halves the candidate list.
What content model do you need? Markdown-only, MDX, or interactive widgets?
Who are the authors? Markdown-comfortable engineers, generalist content writers, or marketers? That decides whether you need CMS integration.
How many languages? One Korean? Pick freely. Five languages? Only first-class i18n tools stay on the list.
Build time and page count? 100 pages is fast in any tool. 10,000 pages means Hugo or Zola.
Need search or AI search? If yes, Mintlify, MkDocs Material, or Starlight have the edge.
Hosting model? Self-host or SaaS-only? Mintlify is SaaS-only.
Can you spend 10,000 dollars per year? If yes, Mintlify is plausible; otherwise stay OSS.
Will you still operate this in five years? Avoid declining tools like Gatsby for new starts.

Common anti-patterns

"Next.js full-stack" for a plain blog. A five-page blog does not need App Router and Server Components. Astro or Hugo is smaller and more stable.
Chasing the AI-shiny tool. Mintlify looks great, but if you cannot commit 10,000 dollars per year, start with MkDocs Material or Starlight and add AI search later.
Ignoring category. Building a giant docs site on a blog-focused SSG (Hugo) or a marketing site on a docs-focused SSG (MkDocs).
Treating multilingual as an afterthought. Write Korean content first, then realize the tool does not do i18n well and rewrite everything. Make i18n a first-class decision factor.
Treating search as an afterthought. Six months pass without an Algolia DocSearch application, while you keep hand-rolling. Decide and integrate early.
Starting new with a declining tool. Starting Gatsby, Hexo, or Jekyll in 2026 is risky on a five-year horizon. Existing sites are fine to keep; new sites should pick tools with momentum.
Underestimating migration. "We'll just swap SSGs" sounds like one week of work and is actually one to three months.
Ignoring build speed. Start with 100 pages, hit 3,000 in a year, watch CI builds creep to 30 minutes and slow every PR. Forecast the page-count curve.

Coming Next

The next post walks through running an internal docs site on MkDocs Material from zero to production — directory layout, multilingual, search, permissions, CI builds, and bolting on AI search.

After that: Astro Starlight versus Docusaurus, head-to-head migration — building the same OSS docs in both tools and recording the quantitative differences.