Dynamic rendering for JavaScript sites with Rendertron and Puppeteer

Our company is engaged in the development, support and maintenance of sites of any complexity. From simple one-page sites to large-scale cluster systems built on micro services. Experience of developers is confirmed by certificates from vendors.
Development and maintenance of all types of websites:
Informational websites or web applications
Business card websites, landing pages, corporate websites, online catalogs, quizzes, promo websites, blogs, news resources, informational portals, forums, aggregators
E-commerce websites or web applications
Online stores, B2B portals, marketplaces, online exchanges, cashback websites, exchanges, dropshipping platforms, product parsers
Business process management web applications
CRM systems, ERP systems, corporate portals, production management systems, information parsers
Electronic service websites or web applications
Classified ads platforms, online schools, online cinemas, website builders, portals for electronic services, video hosting platforms, thematic portals

These are just some of the technical types of websites we work with, and each of them can have its own specific features and functionality, as well as be customized to meet the specific needs and goals of the client.

Our competencies:
Development stages
Latest works
  • image_web-applications_feedme_466_0.webp
    Development of a web application for FEEDME
    1161
  • image_ecommerce_furnoro_435_0.webp
    Development of an online store for the company FURNORO
    1041
  • image_crm_enviok_479_0.webp
    Development of a web application for Enviok
    822
  • image_crm_chasseurs_493_0.webp
    CRM development for Chasseurs
    847
  • image_website-sbh_0.png
    Website development for SBH Partners
    999
  • image_website-_0.png
    Website development for Red Pear
    451

Setting Up Dynamic Rendering for JavaScript Sites (Rendertron/Puppeteer)

Dynamic Rendering is an approach where users get the SPA, but search bots get pre-rendered HTML. Google officially recognizes this as an acceptable workaround when SSR isn't available.

When to Use Dynamic Rendering

  • SPA on React/Vue/Angular without SSR
  • Can't rewrite application for SSR
  • Temporary solution before SSR implementation
  • Specific site sections with indexing problems

Don't use: if SSR/SSG is available — Dynamic Rendering is a second-rate solution with overhead.

Architecture

Request → nginx → User-Agent check
                      ↓
          Bot? → Prerender Service (Puppeteer)
                      ↓
                  HTML snapshot → Bot
                      ↓
          Human? → SPA bundle → Browser

Rendertron: self-hosted prerender

git clone https://github.com/GoogleChrome/rendertron
cd rendertron
npm install
npm run build

# Docker
docker build -t rendertron .
docker run -p 3000:3000 rendertron
# nginx: identify bots and send to Rendertron
map $http_user_agent $prerender_ua {
    ~*(Googlebot|Bingbot|Slurp|DuckDuckBot|Baiduspider|YandexBot|LinkedInBot|
       facebookexternalhit|Twitterbot|WhatsApp|TelegramBot|Slackbot)  1;
    default 0;
}

map $args $prerender_args {
    ~*_escaped_fragment_= 1;
    default 0;
}

map $prerender_ua$prerender_args $prerender {
    "11" 1;
    "10" 1;
    "01" 1;
    default 0;
}

server {
    listen 80;
    server_name company.com;

    location / {
        if ($prerender = 1) {
            rewrite .* /index.html break;
            proxy_pass http://rendertron:3000/render/https://company.com$request_uri;
        }
        try_files $uri /index.html;
    }
}

Custom prerender on Puppeteer

More control over rendering behavior:

// prerender-server.js
const express = require('express')
const puppeteer = require('puppeteer')

const app = express()
let browser = null

async function getBrowser() {
  if (!browser) {
    browser = await puppeteer.launch({
      headless: 'new',
      args: [
        '--no-sandbox',
        '--disable-setuid-sandbox',
        '--disable-dev-shm-usage',
        '--disable-gpu',
      ]
    })
  }
  return browser
}

app.get('/render', async (req, res) => {
  const url = req.query.url
  if (!url) return res.status(400).send('URL required')

  try {
    const browser = await getBrowser()
    const page = await browser.newPage()

    // Block media to speed up
    await page.setRequestInterception(true)
    page.on('request', req => {
      const type = req.resourceType()
      if (['image', 'media', 'font'].includes(type)) {
        req.abort()
      } else {
        req.continue()
      }
    })

    await page.goto(url, {
      waitUntil: 'networkidle0',
      timeout: 30000
    })

    // Wait for specific element (if needed)
    await page.waitForSelector('[data-ssr-ready]', { timeout: 10000 })
      .catch(() => {})  // not fatal if element doesn't exist

    const html = await page.content()
    await page.close()

    // Cache result
    cache.set(url, html, 3600)  // 1 hour

    res.set('Content-Type', 'text/html')
    res.send(html)

  } catch (err) {
    console.error(`Render failed for ${url}:`, err)
    res.status(500).send('Render failed')
  }
})

app.listen(3000)

Caching Prerender Results

Rendering each page takes 1–5 seconds. Cache is mandatory:

const Redis = require('ioredis')
const redis = new Redis({ host: 'redis' })

async function cachedRender(url) {
  const cacheKey = `prerender:${url}`
  const cached = await redis.get(cacheKey)
  if (cached) return cached

  const html = await render(url)
  await redis.setex(cacheKey, 3600, html)
  return html
}

// Invalidate cache on deploy
async function invalidateCache(pathPattern) {
  const keys = await redis.keys(`prerender:*${pathPattern}*`)
  if (keys.length) await redis.del(keys)
}

Middleware in Express/Fastify

// middleware/prerender.js
const botUserAgents = [
  'googlebot', 'bingbot', 'yandexbot', 'baiduspider',
  'facebookexternalhit', 'twitterbot', 'linkedinbot'
]

function isBot(userAgent) {
  return botUserAgents.some(bot =>
    userAgent.toLowerCase().includes(bot)
  )
}

module.exports = async function prerenderMiddleware(req, res, next) {
  const ua = req.headers['user-agent'] || ''

  if (!isBot(ua)) return next()  // human — regular SPA

  try {
    const renderedHtml = await cachedRender(`https://${req.hostname}${req.originalUrl}`)
    res.set('X-Prerendered', 'true')
    res.send(renderedHtml)
  } catch (err) {
    // Fallback: serve regular SPA (better than 500)
    next()
  }
}

Prerender.io as SaaS Alternative

# Using Prerender.io without self-hosted solution
location / {
    if ($prerender = 1) {
        proxy_pass https://service.prerender.io/https://company.com$request_uri;
        proxy_set_header X-Prerender-Token "YOUR_TOKEN";
    }
    try_files $uri /index.html;
}

Monitoring and Debugging

# Check that bot receives rendered HTML
curl -A "Googlebot/2.1 (+http://www.google.com/bot.html)" \
  https://company.com/product/42 | grep -c "product-title"

# Should return number > 0 (elements found in HTML)
// Add marker that page fully rendered
// (for Puppeteer completion condition)
window.prerenderReady = false
// ... after data loaded:
window.prerenderReady = true

// Puppeteer waits for:
await page.waitForFunction(() => window.prerenderReady === true)

Timeline

Setting up Rendertron or custom Puppeteer prerender with nginx and Redis cache — 2–3 business days.