Keep the scraper.
Scroll the rest away.
Your Playwright job, line by line, as the browser farm, proxies, stealth and retries collapse into the call on the right.
import { chromium } from "playwright";
import { newInjectedContext } from "fingerprint-injector";
const proxies = require("./proxies.json");
let attempt = 0;
while (attempt < MAX_RETRIES) {
const proxy = rotate(proxies);
const browser = await chromium.launch({ headless: true, proxy });
const ctx = await newInjectedContext(browser);
const page = await ctx.newPage();
await page.route("**/*", blockResources);
try {
await page.goto(url, { waitUntil: "networkidle" });
await page.waitForSelector(".price", { timeout: 15000 });
await solveCaptcha(page);
const html = await page.content();
const data = parsePrice(html);
await browser.close();
return data;
} catch (e) {
attempt++; await browser.close();
}
}curl "https://api.zenrows.com/v1/" \
-d url=https://store.example/p/123 \
-d js_render=true \
-d premium_proxy=true \
-d wait_for=.price \
-d autoparse=trueBrowser, stealth, proxies and retries are ours to run now.
Delete the
browser farm.
Everything you stand up to make a headless browser look human, and keep it that way, stops being yours.
- A pool of headless Chromium instancesautoscaled, patched, restarted
- Residential & datacenter proxy contractsrotated, scored, rebought
- Fingerprint & stealth patcheschased every Chrome release
- CAPTCHA-solver budgetper-solve, unpredictable
- The 2am page when a site changeson-call, every week
Every line you
babysat is a flag.
Pick a thing your code used to handle. The flag on the right is what replaces all of it.
const ctx = await newInjectedContext(browser);
await solveCaptcha(page);
// + chase every Chrome fingerprint release-d js_render=trueA real browser with a real fingerprint, kept current by us.
Drag the
maintenance away.
The stack you own is on top. Drag the handle right and watch it give way to the one call underneath. Same data out either way.
curl "https://api.zenrows.com/v1/" \
-d url=https://store.example/p/123 \
-d js_render=true \
-d premium_proxy=true \
-d autoparse=true- Browser pool & autoscaling
- Proxy rotation & scoring
- Stealth / fingerprint patches
- CAPTCHA solving
- Retry & backoff logic
- HTML parsing & cleanup
- On-call when a site changes
The output
doesn't change.
Point your existing parser at the response. Same fields, same shape, sourced through one call instead of a fleet.
{
"title": "Aurora 14\" Laptop",
"price": 1299.00,
"currency": "USD",
"in_stock": true,
"rating": 4.6
}Keep the code.
Drop the infrastructure.
The logic that makes your scraper yours stays. The plumbing that made it fragile becomes a line you never touch again.
Migrate one job
this afternoon.
Take the scraper that breaks most, point its URL at Zenrows, and delete the parts you were never paid to maintain.
10,000 credits/month for free, always.