Proxy cho Playwright Puppeteer Selenium 2026

QA và dev automation cần inject proxy vào headless browser cho 3 lý do chính: geo testing (kiểm tra site behavior từ region khác nhau), IP rotation (scrape ToS-compliant không bị rate limit), và anti-bot bypass legitimate (test app phía user thật). Bài này hướng dẫn config proxy IPv4 dedicated TND vào Playwright (Node + Python), Puppeteer, Selenium, kèm code production-ready cho per-context proxy, IP rotation, retry on failure.

1. Vì sao automation cần proxy

• Geo testing: Site hiển thị pricing khác nhau theo country (Netflix, Spotify). Test từ IP US, EU, VN để verify localization.
• A/B test region-specific: Feature flag rollout theo region - automation test từng region.
• Anti-bot bypass legitimate: Site bạn được phép test có Cloudflare/PerimeterX - dùng residential/dedicated IP để mô phỏng user thật.
• Rate limit avoidance: Site limit per-IP, rotate qua nhiều TND IPv4 để scrape bulk hợp pháp.
• Sandbox isolation: Test login flow của customer mà không lộ IP công ty.

2. Playwright Node.js launch option

const { chromium } = require('playwright');

(async () => {
  const browser = await chromium.launch({
    headless: true,
    proxy: {
      server: 'http://23.155.xxx.xxx:2021',
      username: 'us01_S21130_P60',
      password: '••••••••••••••••'
    }
  });
  const page = await browser.newPage();
  await page.goto('https://ipinfo.io');
  console.log(await page.textContent('pre'));
  await browser.close();
})();

SOCKS5:

const browser = await chromium.launch({
  proxy: { server: 'socks5://23.155.xxx.xxx:2020',
           username: 'us01_S21130_P60',
           password: '••••••••' }
});

3. Playwright Python sync + async

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    browser = p.chromium.launch(
        headless=True,
        proxy={
            'server': 'http://23.155.xxx.xxx:2021',
            'username': 'us01_S21130_P60',
            'password': '••••••••'
        }
    )
    page = browser.new_page()
    page.goto('https://ipinfo.io')
    print(page.text_content('pre'))
    browser.close()

Async version:

import asyncio
from playwright.async_api import async_playwright

async def main():
    async with async_playwright() as p:
        browser = await p.chromium.launch(
            proxy={'server': 'http://23.155.xxx.xxx:2021',
                   'username': 'us01_S21130_P60', 'password': '••••••••'}
        )
        page = await browser.new_page()
        await page.goto('https://ipinfo.io')
        print(await page.text_content('pre'))
        await browser.close()

asyncio.run(main())

4. Puppeteer Node.js

const puppeteer = require('puppeteer');

(async () => {
  const browser = await puppeteer.launch({
    headless: true,
    args: ['--proxy-server=http://23.155.xxx.xxx:2021']
  });
  const page = await browser.newPage();
  // Auth qua page.authenticate (Puppeteer không có syntax proxy auth like Playwright)
  await page.authenticate({
    username: 'us01_S21130_P60',
    password: '••••••••'
  });
  await page.goto('https://ipinfo.io');
  console.log(await page.evaluate(() => document.querySelector('pre').innerText));
  await browser.close();
})();

SOCKS5 với Puppeteer:

const browser = await puppeteer.launch({
  args: ['--proxy-server=socks5://23.155.xxx.xxx:2020']
});
// SOCKS5 auth trong Chromium chưa support inline qua args
// Cần extension hoặc tool wrap như proxy-chain

5. Selenium Python ChromeOptions

from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from selenium_authenticated_proxy import SeleniumAuthenticatedProxy

proxy_url = "http://us01_S21130_P60:••••••••@23.155.xxx.xxx:2021"

opts = Options()
opts.add_argument("--headless")

# Cần lib selenium-authenticated-proxy vì Selenium native không support proxy auth
proxy_helper = SeleniumAuthenticatedProxy(proxy_url)
proxy_helper.enrich_chrome_options(opts)

driver = webdriver.Chrome(options=opts)
driver.get("https://ipinfo.io")
print(driver.find_element("css selector", "pre").text)
driver.quit()

Cài lib: pip install selenium-authenticated-proxy

6. Per-context proxy (Playwright)

Playwright cho phép 1 browser instance có nhiều context (giống Chrome profile), mỗi context proxy riêng - efficient hơn launch nhiều browser.

const browser = await chromium.launch({ headless: true });

// Context 1: TND US
const ctx1 = await browser.newContext({
  proxy: { server: 'http://23.155.xxx.xxx:2021',
           username: 'us01_S21130_P60', password: '••••••••' }
});

// Context 2: TND VN
const ctx2 = await browser.newContext({
  proxy: { server: 'http://203.xxx.xxx.xxx:2021',
           username: 'vn01_S21131_P60', password: '••••••••' }
});

const page1 = await ctx1.newPage();
const page2 = await ctx2.newPage();
await page1.goto('https://ipinfo.io');  // → IP US
await page2.goto('https://ipinfo.io');  // → IP VN

7. IP rotation giữa nhiều TND IPv4

const proxies = [
  { server: 'http://23.155.xxx.10:2021', username: '...', password: '...' },
  { server: 'http://23.155.xxx.11:2021', username: '...', password: '...' },
  { server: 'http://23.155.xxx.12:2021', username: '...', password: '...' },
];

let i = 0;
async function scrapeUrl(url) {
  const proxy = proxies[i % proxies.length];
  i++;
  const browser = await chromium.launch({ proxy });
  const page = await browser.newPage();
  await page.goto(url);
  const content = await page.content();
  await browser.close();
  return content;
}

// Loop scrape
const urls = ['https://...page1', 'https://...page2', /*...*/];
for (const url of urls) {
  await scrapeUrl(url);
  await new Promise(r => setTimeout(r, 2000));  // 2s delay polite
}

8. Handle proxy fail + retry

async function scrapeWithRetry(url, maxRetry = 3) {
  for (let attempt = 0; attempt < maxRetry; attempt++) {
    try {
      const proxy = proxies[attempt % proxies.length];
      const browser = await chromium.launch({
        proxy,
        timeout: 15000
      });
      const page = await browser.newPage();
      await page.goto(url, { waitUntil: 'networkidle', timeout: 20000 });
      const content = await page.content();
      await browser.close();
      return content;
    } catch (e) {
      console.warn(`Attempt ${attempt+1} failed: ${e.message}`);
      await new Promise(r => setTimeout(r, 3000));
    }
  }
  throw new Error(`Failed after ${maxRetry} retries: ${url}`);
}

9. Use case thực tế

Use case 1: Geo-test pricing page

SaaS company test pricing page Spotify, Adobe, Microsoft từ 5 country. 5 TND IPv4 (US, UK, DE, FR, JP nếu có) → 5 Playwright context → screenshot + assert pricing localized đúng.

Use case 2: SEO SERP scrape compliant

SEO agency track ranking 100 keyword cho 10 client. Setup: 10 TND IPv4, mỗi keyword 1 invocation, rotate IP, 5s delay/request. Đọc thêm Proxy cho SEO SERP tracking.

Use case 3: E2E test customer flow

App e-commerce test full flow user thật: register → verify email → login → add to cart → checkout. Route automation qua TND US IPv4 để mô phỏng customer US thật, không lộ IP server CI.

Use case 4: Anti-bot bypass legitimate (Cloudflare)

Test site có Cloudflare Bot Management mà bạn được phép test (audit/pentest có hợp đồng). DC proxy thường bị flag, dùng TND IPv4 dedicated với clean reputation history (verify qua AbuseIPDB <5%).

Tổng kết

Playwright + Puppeteer + Selenium đều support proxy qua launch option đơn giản. TND IPv4 dedicated là choice hợp lý cho automation: IP cố định lâu dài (reputation clean), dedicated 1 user, replace miễn phí khi flag. Mỗi tool đều có syntax riêng cho HTTP và SOCKS5 - cần lưu ý Puppeteer chưa native SOCKS5 auth.

Tham khảo thêm:

• Proxy cho Claude Code AI agent
• Proxy cho SEO SERP tracking
• SOCKS5 vs HTTP - chọn protocol
• IPv4 Dedicated vs Residential cho automation