Rate Limiting Your Own CAPTCHA Solving Requests

CaptchaAI handles high concurrency, but you might want to limit yourself. Reasons: cost control, avoiding unintentional spikes from bugs, fair resource sharing across teams, or matching rate limits set by target sites. This guide implements three client-side rate limiting patterns for CaptchaAI.

Why Client-Side Rate Limiting

Scenario	Without limit	With limit
Bug causes infinite solve loop	Burns through balance	Stops at configured cap
Multiple teams share one API key	Uncoordinated spending	Fair allocation per team
Target site bans at > 100 req/min	Accounts get blocked	Stays under threshold
Budget is $50/month	Could exceed in one afternoon	Hard cap enforced

Pattern 1: Token Bucket

A token bucket allows bursts while enforcing an average rate. Tokens replenish at a fixed rate; each request consumes one token.

Python Token Bucket

# token_bucket_solver.py
import os
import time
import threading
import requests

API_KEY = os.environ.get("CAPTCHAAI_KEY", "YOUR_API_KEY")

class TokenBucket:
    """Token bucket rate limiter."""

    def __init__(self, rate, capacity):
        """
        rate: tokens added per second
        capacity: max tokens (burst size)
        """
        self.rate = rate
        self.capacity = capacity
        self.tokens = capacity
        self.last_refill = time.monotonic()
        self.lock = threading.Lock()

    def acquire(self, timeout=30):
        """Wait for a token. Returns True if acquired, False on timeout."""
        deadline = time.monotonic() + timeout
        while True:
            with self.lock:
                self._refill()
                if self.tokens >= 1:
                    self.tokens -= 1
                    return True

            if time.monotonic() >= deadline:
                return False
            time.sleep(0.1)

    def _refill(self):
        now = time.monotonic()
        elapsed = now - self.last_refill
        self.tokens = min(self.capacity, self.tokens + elapsed * self.rate)
        self.last_refill = now

# Allow 10 solves/minute with burst of 5
limiter = TokenBucket(rate=10/60, capacity=5)

def solve_rate_limited(sitekey, pageurl):
    """Solve with rate limiting."""
    if not limiter.acquire(timeout=60):
        raise Exception("Rate limit: could not acquire token within 60s")

    session = requests.Session()
    resp = session.get("https://ocr.captchaai.com/in.php", params={
        "key": API_KEY,
        "method": "userrecaptcha",
        "googlekey": sitekey,
        "pageurl": pageurl,
        "json": "1",
    })
    result = resp.json()

    if result.get("status") != 1:
        raise Exception(f"Submit failed: {result.get('request')}")

    task_id = result["request"]
    time.sleep(15)

    for _ in range(25):
        poll = session.get("https://ocr.captchaai.com/res.php", params={
            "key": API_KEY, "action": "get",
            "id": task_id, "json": "1",
        })
        poll_result = poll.json()
        if poll_result.get("status") == 1:
            return poll_result["request"]
        if poll_result.get("request") != "CAPCHA_NOT_READY":
            raise Exception(f"Error: {poll_result.get('request')}")
        time.sleep(5)

    raise Exception("Timeout")

Pattern 2: Sliding Window Counter

Tracks the number of requests within a fixed time window. Simpler than token bucket but no burst control.

JavaScript Sliding Window

// sliding_window_solver.js
const axios = require('axios');

const API_KEY = process.env.CAPTCHAAI_KEY || 'YOUR_API_KEY';

class SlidingWindowLimiter {
  constructor(maxRequests, windowMs) {
    this.maxRequests = maxRequests;
    this.windowMs = windowMs;
    this.timestamps = [];
  }

  async acquire(timeoutMs = 60000) {
    const deadline = Date.now() + timeoutMs;

    while (Date.now() < deadline) {
      // Remove expired timestamps
      const cutoff = Date.now() - this.windowMs;
      this.timestamps = this.timestamps.filter(t => t > cutoff);

      if (this.timestamps.length < this.maxRequests) {
        this.timestamps.push(Date.now());
        return true;
      }

      // Wait until the oldest request exits the window
      const waitMs = Math.min(
        this.timestamps[0] + this.windowMs - Date.now() + 10,
        deadline - Date.now()
      );
      if (waitMs > 0) await new Promise(r => setTimeout(r, waitMs));
    }
    return false;
  }
}

// Allow 20 solves per 5 minutes
const limiter = new SlidingWindowLimiter(20, 5 * 60 * 1000);

async function solveRateLimited(sitekey, pageurl) {
  const acquired = await limiter.acquire(60000);
  if (!acquired) throw new Error('Rate limit exceeded');

  const submit = await axios.get('https://ocr.captchaai.com/in.php', {
    params: {
      key: API_KEY, method: 'userrecaptcha',
      googlekey: sitekey, pageurl, json: '1',
    },
  });

  if (submit.data.status !== 1) throw new Error(submit.data.request);
  await new Promise(r => setTimeout(r, 15000));

  for (let i = 0; i < 25; i++) {
    const poll = await axios.get('https://ocr.captchaai.com/res.php', {
      params: { key: API_KEY, action: 'get', id: submit.data.request, json: '1' },
    });
    if (poll.data.status === 1) return poll.data.request;
    if (poll.data.request !== 'CAPCHA_NOT_READY') throw new Error(poll.data.request);
    await new Promise(r => setTimeout(r, 5000));
  }
  throw new Error('Timeout');
}

Pattern 3: Budget-Based Limiter

Set a daily cost budget and stop when it's reached:

# budget_limiter.py
import os
import time
from datetime import date

class BudgetLimiter:
    """Limit daily CAPTCHA spending."""

    def __init__(self, daily_budget, cost_per_solve=0.003):
        self.daily_budget = daily_budget
        self.cost_per_solve = cost_per_solve
        self.daily_spend = 0.0
        self.current_date = date.today()

    def can_solve(self):
        """Check if budget allows another solve."""
        if date.today() != self.current_date:
            self.daily_spend = 0.0
            self.current_date = date.today()

        return self.daily_spend + self.cost_per_solve <= self.daily_budget

    def record_solve(self):
        """Record a successful solve against the budget."""
        self.daily_spend += self.cost_per_solve

    @property
    def remaining_budget(self):
        return max(0, self.daily_budget - self.daily_spend)

    @property
    def remaining_solves(self):
        return int(self.remaining_budget / self.cost_per_solve)

# $5/day budget
budget = BudgetLimiter(daily_budget=5.00, cost_per_solve=0.003)

def solve_with_budget(sitekey, pageurl):
    if not budget.can_solve():
        raise Exception(
            f"Daily budget exhausted. Remaining: ${budget.remaining_budget:.2f}"
        )

    # ... solve logic ...
    token = "..."  # actual solve
    budget.record_solve()
    return token

Choosing a Pattern

Pattern	Best For	Complexity
Token bucket	Smooth rate control with burst allowance	Medium
Sliding window	Simple request count per time window	Low
Budget limiter	Cost control per day/week/month	Low
Combined (rate + budget)	Production systems	Medium

Troubleshooting

Issue	Cause	Fix
All requests queued, none executing	Rate too low for demand	Increase rate or window size
Budget limiter resets mid-day	System clock change or restart	Persist daily spend to file or database
Token bucket drains in burst	Capacity too small for workflow	Increase capacity parameter
Rate limiter blocks polling requests	Limiter applied to polls too	Only limit submit requests, not polls

FAQ

Should I rate limit the submit request or the poll request?

Rate limit only the submit (in.php) request. Polling (res.php) should run freely since it doesn't create new tasks or cost money.

Use Redis-backed rate limiting. Libraries like redis-rate-limiter (Python) or rate-limiter-flexible (Node.js) support distributed rate limiting with atomic operations.