SEO SSR and prerendering

Lovable apps are Vite + React SPAs. Crawlers and link-preview bots see an empty <div id="root"></div> until JS runs. Public marketing pages must ship populated HTML — Google does run JS, but indexing is slower and less reliable than HTML-first; Twitter/LinkedIn/Slack/Discord bots do not run JS at all.

Companion skills: seo-meta-tags-spa, seo-landing-page, tanstack-prefetch-and-hydration, programmatic-seo-pages.

When to use what

Start with prerendering; move to SSR only when content changes too fast for a rebuild.

Build-time prerender (vite-ssg pattern)

// vite.config.ts
import { defineConfig } from "vite";
import react from "@vitejs/plugin-react";

export default defineConfig({
  plugins: [react()],
  // vite-ssg config in package.json or its own file
});

// src/main.tsx
import { ViteSSG } from "vite-ssg";
import App from "./App";
import { routes } from "./routes";

export const createApp = ViteSSG(App, { routes }, ({ app, router, isClient, initialState }) => {
  // hydrate or set up per-app state
});

For each public route, expose an includedRoutes generator that lists URLs (e.g. one per blog post slug, one per location).

Prerender service (prerender.io / Browserless)

If you can't change the build pipeline:

1. Add a serverless function or middleware that detects bot User-Agents and proxies to prerender.io.
2. Prerender.io caches the rendered HTML and serves it to crawlers.

This is fine as a stopgap but slower and not free at scale.

react-snap (alternative, low-effort)

npm i -D react-snap

{
  "scripts": { "postbuild": "react-snap" }
}

Edit index.html to use hydrateRoot instead of createRoot so the prerendered HTML hydrates cleanly.

Caveats: react-snap is unmaintained but still works for static SPAs.

Hydration with data

Don't prerender a "loading…" shell. Either:

• Fetch data at build time and pass it through, or
• Use tanstack-prefetch-and-hydration to embed a dehydrated cache and hydrate on the client without a refetch flash.

Routes to prerender — checklist

• / (home)
• /pricing
• /blog and every /blog/[slug]
• /about, /contact, /changelog, /legal/*
• Programmatic templates (/integrations/[tool], /locations/[city]) — see programmatic-seo-pages
• 404 page (/404.html)

Do not prerender authenticated routes (/app, /dashboard, /settings) — they should redirect to login.

Verify

For every public URL, raw HTML must include the title, meta, and main content:

curl -A "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)" \
  https://your-domain.com/pricing | grep -iE '<title|<h1|og:title'

Also check Twitter card / Open Graph rendering: https://cards-dev.twitter.com/validator, https://developers.facebook.com/tools/debug/.

Avoid

• Shipping a Lovable marketing site as a pure SPA and hoping Google figures it out.
• Prerendering authenticated routes and leaking dashboard chrome to crawlers.
• Different HTML for bots vs users (cloaking) — Google penalizes.
• Forgetting hydrateRoot after introducing prerender — causes a re-render flash.

Checklist

• Every public route ships with title, meta, h1, and primary copy in raw HTML.
• OG image and Twitter card validated in third-party debuggers.
• Auth routes excluded from prerender.
• Sitemap and robots.txt in sync with prerendered URL list (seo-sitemap-generation).