A Puppeteer alternative for screenshots & PDFs

Puppeteer and Playwright are great libraries. Running them in production just to make screenshots or PDFs costs more than it looks.

What self-hosting headless Chrome involves

• A ~300MB browser in your image, plus system fonts (CJK, emoji) so pages don't render as tofu.
• Memory spikes and zombie processes. You need a browser pool, timeouts and restarts.
• Serverless limits: bundle size and cold starts on Lambda or Vercel need special Chromium builds.
• Security: rendering user-supplied URLs from inside your network is an SSRF risk unless you lock down egress.

The API version

// Before: puppeteer
const browser = await puppeteer.launch();
const page = await browser.newPage();
await page.goto(url, { waitUntil: "networkidle0" });
const png = await page.screenshot({ fullPage: true });
await browser.close();

// After: one HTTP call
const r = await fetch("https://shotmint.creasework.com/v1/screenshot?" + new URLSearchParams({ url, full_page: "true", wait_until: "networkidle" }),
  { headers: { "X-API-Key": process.env.SHOTMINT_KEY } });
const png = Buffer.from(await r.arrayBuffer());

Parameter mapping

Puppeteer	Shotmint
page.setViewport({width, height, deviceScaleFactor})	width, height, device_scale_factor
screenshot({fullPage: true})	full_page=true
(await page.$(sel)).screenshot()	selector=sel
emulateMediaFeatures([{name:'prefers-color-scheme',value:'dark'}])	dark_mode=true
page.pdf({format, landscape, margin, printBackground, scale})	/v1/pdf with format, landscape, margin, print_background, scale
page.setContent(html)	html in the POST body

When you should keep Puppeteer

Be honest with yourself here. If you need multi-step flows (log in, click, fill forms), you render millions of pages a month, or pages are only reachable inside your private network, self-hosting is the right call. An API pays off when rendering is a side feature you'd rather not operate.