粘性会话与旋转会话：验证码的最佳代理策略

粘性会话在设定的持续时间内保留相同的 IP。轮换会话为每个请求分配一个新的 IP。选择错误的验证码会提高验证码率并浪费 API 积分。

每种模式如何运作

STICKY SESSION:
  Request 1 ──▶ IP: 192.168.1.50 ──▶ site.com/login
  Request 2 ──▶ IP: 192.168.1.50 ──▶ site.com/dashboard
  Request 3 ──▶ IP: 192.168.1.50 ──▶ site.com/account
  (Same IP for all requests in the session window)

ROTATING SESSION:
  Request 1 ──▶ IP: 192.168.1.50 ──▶ site.com/page1
  Request 2 ──▶ IP: 10.0.0.77   ──▶ site.com/page2
  Request 3 ──▶ IP: 172.16.5.22 ──▶ site.com/page3
  (Different IP per request)

头对头比较

当粘性会话获胜时

多步骤工作流程

CAPTCHA 根据请求的 IP 验证令牌。如果您的 IP 在解决和提交之间发生变化，则令牌可能会被拒绝。

import requests
import time

CAPTCHAAI_KEY = "YOUR_API_KEY"
CAPTCHAAI_URL = "https://ocr.captchaai.com"

# Sticky session: same IP for the entire login flow
STICKY_PROXY = {
    "http": "http://user-session-abc123:pass@proxy.example.com:5000",
    "https": "http://user-session-abc123:pass@proxy.example.com:5000",
}

session = requests.Session()
session.proxies = STICKY_PROXY
session.headers.update({
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
    "AppleWebKit/537.36 Chrome/126.0.0.0 Safari/537.36",
})


def login_with_captcha(url, sitekey, username, password):
    """Full login flow — MUST use same IP throughout."""
    # Step 1: Load login page (IP: 192.168.1.50)
    session.get(url)

    # Step 2: Solve CAPTCHA (solved against IP: 192.168.1.50)
    token = solve_recaptcha(sitekey, url)

    # Step 3: Submit login (MUST be IP: 192.168.1.50)
    resp = session.post(url, data={
        "username": username,
        "password": password,
        "g-recaptcha-response": token,
    })

    return resp.status_code == 200


def solve_recaptcha(sitekey, pageurl):
    resp = requests.post(f"{CAPTCHAAI_URL}/in.php", data={
        "key": CAPTCHAAI_KEY,
        "method": "userrecaptcha",
        "googlekey": sitekey,
        "pageurl": pageurl,
        "json": 1,
    })
    task_id = resp.json()["request"]

    for _ in range(60):
        time.sleep(5)
        result = requests.get(f"{CAPTCHAAI_URL}/res.php", params={
            "key": CAPTCHAAI_KEY, "action": "get",
            "id": task_id, "json": 1,
        })
        data = result.json()
        if data["request"] != "CAPCHA_NOT_READY":
            return data["request"]

    raise TimeoutError("CAPTCHA solve timeout")

为什么 IP 一致性很重要

Sticky session:

  1. Browser loads page    → IP: 5.5.5.5
  2. CAPTCHA solved        → Token bound to session
  3. Form submitted        → IP: 5.5.5.5 ✅ Token accepted

Rotating session:

  1. Browser loads page    → IP: 5.5.5.5
  2. CAPTCHA solved        → Token bound to session
  3. Form submitted        → IP: 9.9.9.9 ❌ Token may be rejected

当轮流会议获胜时

批量页面抓取

每个页面都是独立的。轮换 IP 可将请求分散到更大的池中，从而减少每个 IP 的检测。

import concurrent.futures

ROTATING_PROXY = {
    "http": "http://user:pass@rotating.proxy.example.com:5000",
    "https": "http://user:pass@rotating.proxy.example.com:5000",
}


def scrape_page(url):
    """Each request gets a fresh IP automatically."""
    resp = requests.get(
        url,
        proxies=ROTATING_PROXY,
        headers={"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64)"},
        timeout=30,
    )
    if resp.status_code == 200:
        return url, resp.text
    return url, None


urls = [f"https://example.com/products?page={i}" for i in range(1, 501)]

with concurrent.futures.ThreadPoolExecutor(max_workers=10) as pool:
    results = list(pool.map(scrape_page, urls))
    success = sum(1 for _, html in results if html)
    print(f"Scraped: {success}/{len(urls)}")

搜索引擎抓取

搜索引擎跟踪每个 IP 的查询量。轮换避免速率限制：

def scrape_search(query, pages=10):
    results = []
    for page in range(pages):
        # Each request → new IP
        resp = requests.get(
            "https://search-engine.example.com/search",
            params={"q": query, "start": page * 10},
            proxies=ROTATING_PROXY,
            timeout=20,
        )
        results.append(resp.text)
        time.sleep(2)
    return results

混合策略

在同一管道中使用两种模式：

class HybridProxyManager:
    """Sticky for multi-step flows, rotating for single requests."""

    def __init__(self, provider_host, username, password, port=5000):
        self.host = provider_host
        self.username = username
        self.password = password
        self.port = port

    def get_rotating_proxy(self):
        return {
            "http": f"http://{self.username}:{self.password}@{self.host}:{self.port}",
            "https": f"http://{self.username}:{self.password}@{self.host}:{self.port}",
        }

    def get_sticky_proxy(self, session_id, duration_min=10):
        sticky_user = f"{self.username}-session-{session_id}-ttl-{duration_min}"
        return {
            "http": f"http://{sticky_user}:{self.password}@{self.host}:{self.port}",
            "https": f"http://{sticky_user}:{self.password}@{self.host}:{self.port}",
        }


proxy_mgr = HybridProxyManager("proxy.example.com", "user", "pass")


def scrape_with_auto_strategy(url, needs_login=False):
    if needs_login:
        # Multi-step → sticky
        import uuid
        session_id = uuid.uuid4().hex[:8]
        proxy = proxy_mgr.get_sticky_proxy(session_id)
    else:
        # Single page → rotating
        proxy = proxy_mgr.get_rotating_proxy()

    return requests.get(url, proxies=proxy, timeout=30)

Node.js 实现

const axios = require("axios");

const CAPTCHAAI_KEY = "YOUR_API_KEY";
const PROXY_HOST = "proxy.example.com";
const PROXY_PORT = 5000;

function getProxy(mode, sessionId = null) {
  const user =
    mode === "sticky" ? `user-session-${sessionId}` : "user";

  return {
    proxy: {
      host: PROXY_HOST,
      port: PROXY_PORT,
      auth: { username: user, password: "pass" },
    },
  };
}

// Rotating: bulk scrape
async function scrapePages(urls) {
  const results = [];
  for (const url of urls) {
    const config = getProxy("rotating");
    const resp = await axios.get(url, { ...config, timeout: 30000 });
    results.push({ url, data: resp.data });
  }
  return results;
}

// Sticky: login flow
async function loginFlow(loginUrl, sitekey, credentials) {
  const sessionId = Date.now().toString(36);
  const config = getProxy("sticky", sessionId);

  // Step 1: Load page
  await axios.get(loginUrl, config);

  // Step 2: Solve CAPTCHA
  const submitResp = await axios.post(
    "https://ocr.captchaai.com/in.php",
    null,
    {
      params: {
        key: CAPTCHAAI_KEY,
        method: "userrecaptcha",
        googlekey: sitekey,
        pageurl: loginUrl,
        json: 1,
      },
    }
  );
  const taskId = submitResp.data.request;

  let token;
  for (let i = 0; i < 60; i++) {
    await new Promise((r) => setTimeout(r, 5000));
    const res = await axios.get("https://ocr.captchaai.com/res.php", {
      params: { key: CAPTCHAAI_KEY, action: "get", id: taskId, json: 1 },
    });
    if (res.data.request !== "CAPCHA_NOT_READY") {
      token = res.data.request;
      break;
    }
  }

  // Step 3: Submit with same IP
  return axios.post(
    loginUrl,
    {
      ...credentials,
      "g-recaptcha-response": token,
    },
    config
  );
}

决策矩阵

粘性会话持续时间

经验法则： 将粘性持续时间设置为预期流动时间的 2 倍。

故障排除

常问问题

我可以在一个项目中混合粘性和旋转吗？

是的。使用混合方法——粘性用于身份验证流程，轮换用于数据收集。大多数代理提供商在同一计划上支持这两种模式。

粘性会话能保证同一个IP吗？

大多数提供商都会在 TTL 窗口内保证它。如果 IP 变得不可用（运营商重新分配、池轮换），则会分配新的 IP。

哪种模式更便宜？

价格相同——无论模式如何，您都按 GB 或每个请求付费。成本差异来自于验证码费率：错误模式 → 更多验证码 → 更高的 CaptchaAI 花费。

如何设置与代理提供商的粘性会话？

大多数在用户名中使用会话 ID：user-session-abc123。检查您的提供商的文档以了解确切的格式。我们涵盖明亮的数据,智能代理， 和氧化实验室在专门的指南中。

相关指南

• 轮换自有服务器基础设施
• 代理质量影响解决率
• 浏览器会话持久性

将您的代理策略与您的工作流程相匹配 -获取您的 CaptchaAI 密钥在任何会话模式下处理验证码。