Manejar bloqueos de Cloudflare: guía de acceso legítimo

How Cloudflare Detection Works

Cloudflare is the most widely deployed anti-bot service, protecting over 20% of all websites. Understanding how it detects automated traffic is essential for anyone building legitimate scraping tools. Cloudflare uses a multi-layered detection pipeline:

1. IP reputation scoring: Cloudflare maintains a global threat intelligence database. Datacenter IPs, known VPN ranges, and previously flagged addresses receive higher risk scores.
2. TLS fingerprinting: Cloudflare analyzes TLS ClientHello messages to determine if the connecting client matches its claimed identity.
3. Browser fingerprinting: JavaScript challenges probe canvas, WebGL, navigator properties, and dozens of other signals.
4. JavaScript challenges: Cloudflare serves JavaScript that must execute correctly in a real browser environment.
5. Behavioral analysis: Request timing, navigation patterns, mouse movements, and interaction signals are analyzed.
6. Machine learning models: All signals are fed into ML models that continuously adapt to new automation patterns.

For a broader overview, see our comprehensive guide to anti-bot detection systems.

Cloudflare Protection Tiers

Tier	Detection Methods	Difficulty Level	Typical Sites
Basic (Free)	IP reputation, basic JS challenge	Low	Small blogs, personal sites
Pro	+ WAF rules, rate limiting	Medium	Medium businesses, SaaS
Business	+ Advanced Bot Management	High	E-commerce, enterprise sites
Enterprise	+ ML-powered bot scoring, behavioral analysis	Very High	Major retailers, financial services

Ethical Framework for Accessing Cloudflare-Protected Sites

Before implementing any technical approach, establish clear ethical boundaries:

• Check for APIs first: Many Cloudflare-protected sites offer official APIs for data access. Always prefer these.
• Respect robots.txt: If the site explicitly disallows scraping specific paths, honor those directives.
• Review terms of service: Understand what the site permits regarding automated access.
• Access only public data: Never attempt to bypass authentication or access private data.
• Minimize server impact: Use reasonable request rates and do not overload the target server.
• Consider data licensing: For commercial use cases, explore data licensing agreements.

The techniques in this guide are designed for legitimate access to publicly available data. They should never be used to circumvent security protections for unauthorized access, credential theft, or denial-of-service attacks.

Strategy 1: Residential Proxies with Clean IPs

The most effective first step is ensuring your IP addresses have clean reputations. Cloudflare's IP scoring heavily penalizes datacenter and VPN IPs.

# Python: Using residential proxies for Cloudflare-protected sites
from curl_cffi import requests as curl_requests
response = curl_requests.get(
    "https://cloudflare-protected-site.com",
    impersonate="chrome",
    proxies={
        "http": "http://USERNAME:PASSWORD@gate.proxyhat.com:8080",
        "https": "http://USERNAME:PASSWORD@gate.proxyhat.com:8080"
    },
    timeout=30
)
if response.status_code == 200:
    print("Access granted")
elif response.status_code == 403:
    print("Blocked — may need additional measures")
elif response.status_code == 503:
    print("Cloudflare challenge page — need browser execution")

ProxyHat's residential proxies provide IPs classified as genuine residential addresses in Cloudflare's database, bypassing the IP reputation layer. See our comparison of residential proxies vs VPNs for why VPN IPs fail against Cloudflare.

Strategy 2: Browser-Grade TLS Fingerprints

Cloudflare checks JA3/JA4 TLS fingerprints to identify the connecting client. Python's requests library, Go's net/http, and Node.js's default clients all produce non-browser TLS signatures that Cloudflare flags.

Client	Cloudflare Result	Why
Python requests	Blocked or challenged	OpenSSL TLS fingerprint is non-browser
curl_cffi (impersonate="chrome")	Usually passes	Mimics Chrome BoringSSL fingerprint
Headless Chrome (Puppeteer/Playwright)	Usually passes	Real BoringSSL TLS stack
Go net/http	Blocked or challenged	Go crypto/tls fingerprint is distinctive
Go with uTLS (Chrome hello)	Usually passes	Mimics Chrome fingerprint

Strategy 3: Handling JavaScript Challenges

Cloudflare's JavaScript challenges require a real browser environment to solve. There are two approaches:

Approach A: Headless Browser

// Node.js: Playwright with stealth for Cloudflare challenges
const { chromium } = require('playwright');
async function accessCloudflare(url) {
  const browser = await chromium.launch({
    proxy: {
      server: 'http://gate.proxyhat.com:8080',
      username: 'USERNAME',
      password: 'PASSWORD'
    }
  });
  const context = await browser.newContext({
    locale: 'en-US',
    timezoneId: 'America/New_York',
    viewport: { width: 1920, height: 1080 }
  });
  const page = await context.newPage();
  // Navigate and wait for Cloudflare challenge to resolve
  await page.goto(url, { waitUntil: 'networkidle', timeout: 60000 });
  // Cloudflare challenges typically redirect after completion
  // Wait for the actual content to load
  await page.waitForSelector('body', { timeout: 30000 });
  // Check if we passed the challenge
  const title = await page.title();
  if (title.includes('Just a moment') || title.includes('Attention Required')) {
    // Challenge not yet resolved — wait longer
    await page.waitForNavigation({ waitUntil: 'networkidle', timeout: 30000 });
  }
  const content = await page.content();
  await browser.close();
  return content;
}

Approach B: Cookie Extraction and Reuse

Solve the challenge once in a headless browser, extract the cookies (especially cf_clearance), then reuse them in a lightweight HTTP client:

// Node.js: Extract Cloudflare cookies for reuse
const { chromium } = require('playwright');
async function extractCfCookies(url) {
  const browser = await chromium.launch({
    proxy: {
      server: 'http://gate.proxyhat.com:8080',
      username: 'USERNAME-session-cf1',
      password: 'PASSWORD'
    }
  });
  const context = await browser.newContext({
    locale: 'en-US',
    timezoneId: 'America/New_York',
  });
  const page = await context.newPage();
  await page.goto(url, { waitUntil: 'networkidle', timeout: 60000 });
  // Wait for challenge resolution
  await page.waitForTimeout(10000);
  // Extract cookies
  const cookies = await context.cookies();
  const cfClearance = cookies.find(c => c.name === 'cf_clearance');
  const userAgent = await page.evaluate(() => navigator.userAgent);
  await browser.close();
  return { cookies, userAgent, cfClearance };
}
// Reuse cookies with got-scraping (same proxy session!)
import { gotScraping } from 'got-scraping';
const { cookies, userAgent } = await extractCfCookies('https://example.com');
const cookieString = cookies.map(c => `${c.name}=${c.value}`).join('; ');
const response = await gotScraping({
  url: 'https://example.com/api/data',
  proxyUrl: 'http://USERNAME-session-cf1:PASSWORD@gate.proxyhat.com:8080',
  headers: {
    'Cookie': cookieString,
    'User-Agent': userAgent,  // Must match the browser that solved the challenge
  }
});

Important: The cf_clearance cookie is bound to the IP address and user-agent that solved the challenge. You must use the same proxy session (sticky IP) and identical user-agent when reusing it.

Strategy 4: Request Pattern Optimization

Cloudflare's behavioral analysis flags non-human request patterns. Follow these patterns for legitimate access:

Realistic Navigation Flow

# Python: Realistic navigation pattern
from curl_cffi import requests as curl_requests
import time
import random
session = curl_requests.Session(impersonate="chrome")
session.proxies = {
    "http": "http://USERNAME:PASSWORD@gate.proxyhat.com:8080",
    "https": "http://USERNAME:PASSWORD@gate.proxyhat.com:8080"
}
# Step 1: Visit homepage first
home = session.get("https://example.com")
time.sleep(random.uniform(2.0, 4.0))
# Step 2: Navigate to category (with Referer)
category = session.get(
    "https://example.com/products",
    headers={"Referer": "https://example.com"}
)
time.sleep(random.uniform(1.5, 3.5))
# Step 3: Browse items (with proper Referer chain)
for item_url in item_urls[:20]:
    item = session.get(
        item_url,
        headers={"Referer": "https://example.com/products"}
    )
    time.sleep(random.uniform(1.0, 3.0))

Rate Limiting Guidelines

Cloudflare Tier	Safe Request Rate	Delay Between Requests
Basic/Free	20-30 req/min	2-3 seconds
Pro	10-20 req/min	3-6 seconds
Business	5-10 req/min	6-12 seconds
Enterprise	2-5 req/min	12-30 seconds

Strategy 5: Handling Common Cloudflare Responses

Status Code	Meaning	Action
200	Success	Parse content normally
403	Forbidden — IP or fingerprint blocked	Rotate to a new IP, check TLS fingerprint
429	Rate limited	Back off exponentially, reduce request rate
503	JavaScript challenge	Use headless browser to solve
520-527	Cloudflare server errors	Retry after delay — origin server issue

# Python: Response handling with retry logic
import time
import random
def cloudflare_resilient_request(session, url, max_retries=3):
    for attempt in range(max_retries):
        try:
            response = session.get(url, timeout=30)
            if response.status_code == 200:
                return response
            if response.status_code == 403:
                # IP flagged — rotate session
                print(f"403 on attempt {attempt + 1} — rotating IP")
                session = create_new_session()
                time.sleep(random.uniform(5, 10))
                continue
            if response.status_code == 429:
                # Rate limited — exponential backoff
                wait = (2 ** attempt) * 5 + random.uniform(0, 5)
                print(f"429 — waiting {wait:.1f}s")
                time.sleep(wait)
                continue
            if response.status_code == 503:
                # JS challenge — need headless browser
                print("503 — JavaScript challenge detected")
                return None  # Escalate to browser-based approach
        except Exception as e:
            print(f"Error: {e}")
            time.sleep(random.uniform(2, 5))
    return None

Complete Multi-Layer Approach

The most reliable strategy combines all layers:

1. Residential proxies: ProxyHat residential IPs for clean IP reputation.
2. Browser-grade TLS: curl_cffi or headless browser for correct fingerprints.
3. Consistent headers: Complete header sets matching the claimed browser.
4. Natural timing: Randomized delays following human browsing patterns.
5. Cookie management: Accept and maintain cookies throughout sessions.
6. Referer chains: Proper navigation flow from homepage to target pages.

For comprehensive detection reduction strategies, see our complete anti-detection guide. For proxy integration across programming languages, see our guides for Python, Node.js, and Go.

When Not to Scrape

Recognize situations where scraping is not the right approach:

• The site has a public API: Always use official APIs when available.
• The data is behind authentication: Accessing login-protected data via scraping is typically a ToS violation.
• The site explicitly prohibits scraping: Respect clear prohibitions in the ToS.
• Data licensing is available: For commercial use, purchasing data licenses is often more reliable and legal.
• The content is copyrighted: Scraping copyrighted content for redistribution raises legal concerns.

Refer to ProxyHat's documentation for responsible usage guidelines and terms of service.

Preguntas frecuentes