BLS CAPTCHA: Instructions Codes and Solving Guide

BLS CAPTCHAs present a 3×3 grid of images with a numeric instruction code. The user must select cells that match the instruction. CaptchaAI solves BLS CAPTCHAs with 100% accuracy — this guide covers how to extract the grid, read the instruction code, and submit everything to the API.

How BLS CAPTCHAs work

A BLS CAPTCHA displays:

A 3×3 grid containing 9 image cells
A numeric instruction code (e.g., 664, 123, 546) specifying which cells to select
Cells are numbered left-to-right, top-to-bottom:

1  2  3
4  5  6
7  8  9

The instruction code tells the solver which pattern to look for. The response is an array of cell indices (1–9) that match.

Step 1: Extract grid images and instruction code

Python (Selenium)

import base64
from selenium import webdriver
from selenium.webdriver.common.by import By

driver = webdriver.Chrome()
driver.get("https://example.com/bls-protected-page")

# Find the grid container
grid_cells = driver.find_elements(By.CSS_SELECTOR, ".captcha-grid img")

images = []
for cell in grid_cells:
    src = cell.get_attribute("src")
    if src.startswith("data:image"):
        images.append(src)
    else:
        # Download and convert to base64
        import requests
        img_data = requests.get(src).content
        b64 = base64.b64encode(img_data).decode()
        images.append(f"data:image/png;base64,{b64}")

# Extract the instruction code
instruction_el = driver.find_element(By.CSS_SELECTOR, ".captcha-instruction")
instruction_code = instruction_el.text.strip()
# e.g., "664" or parsed from "Select all boxes with number 664"

import re
code_match = re.search(r'(\d{3,})', instruction_code)
instruction = code_match.group(1) if code_match else instruction_code

print(f"Instruction: {instruction}")
print(f"Images extracted: {len(images)}")

JavaScript (Puppeteer)

const puppeteer = require('puppeteer');

const browser = await puppeteer.launch();
const page = await browser.newPage();
await page.goto('https://example.com/bls-protected-page');

// Extract grid images as base64
const images = await page.evaluate(() => {
  const cells = document.querySelectorAll('.captcha-grid img');
  return Array.from(cells).map(img => {
    const canvas = document.createElement('canvas');
    canvas.width = img.naturalWidth;
    canvas.height = img.naturalHeight;
    canvas.getContext('2d').drawImage(img, 0, 0);
    return canvas.toDataURL('image/png');
  });
});

// Extract instruction code
const instruction = await page.evaluate(() => {
  const el = document.querySelector('.captcha-instruction');
  const match = el.textContent.match(/(\d{3,})/);
  return match ? match[1] : el.textContent.trim();
});

console.log(`Instruction: ${instruction}, Images: ${images.length}`);

Step 2: Submit to CaptchaAI

The BLS solver requires method=bls, the instructions code, and all 9 images as image_base64_1 through image_base64_9.

Python

import requests
import time
import json

API_KEY = "YOUR_API_KEY"

# Prepare submission data
data = {
    "key": API_KEY,
    "method": "bls",
    "instructions": instruction,
    "json": "1",
}

# Add all 9 images
files = {}
for i, img in enumerate(images):
    files[f"image_base64_{i+1}"] = (None, img)

# Submit
resp = requests.post(
    "https://ocr.captchaai.com/in.php",
    data=data,
    files=files
).json()

if resp["status"] != 1:
    raise Exception(f"Submit error: {resp['request']}")

task_id = resp["request"]
print(f"Task ID: {task_id}")

# Poll for result
for _ in range(20):
    time.sleep(5)
    result = requests.get("https://ocr.captchaai.com/res.php", params={
        "key": API_KEY, "action": "get", "id": task_id, "json": "1"
    }).json()

    if result["status"] == 1:
        solution = json.loads(result["request"])
        print(f"Selected cells: {solution}")  # e.g., [1, 4, 7, 8]
        break
    if result["request"] != "CAPCHA_NOT_READY":
        raise Exception(f"Error: {result['request']}")

JavaScript

const axios = require('axios');
const FormData = require('form-data');

const form = new FormData();
form.append('key', 'YOUR_API_KEY');
form.append('method', 'bls');
form.append('instructions', instruction);
form.append('json', '1');

images.forEach((img, i) => {
  form.append(`image_base64_${i + 1}`, img);
});

const submit = await axios.post('https://ocr.captchaai.com/in.php', form, {
  headers: form.getHeaders(),
});
const taskId = submit.data.request;

// Poll
let solution = null;
for (let i = 0; i < 20; i++) {
  await new Promise(r => setTimeout(r, 5000));
  const poll = await axios.get('https://ocr.captchaai.com/res.php', {
    params: { key: 'YOUR_API_KEY', action: 'get', id: taskId, json: 1 }
  });
  if (poll.data.status === 1) {
    solution = JSON.parse(poll.data.request);
    break;
  }
}
console.log('Selected cells:', solution); // e.g., [2, 4, 7]

Step 3: Click the solved cells

# Selenium — click the cells returned by CaptchaAI
grid_cells = driver.find_elements(By.CSS_SELECTOR, ".captcha-grid .cell")

for cell_index in solution:
    # cell_index is 1-based
    grid_cells[cell_index - 1].click()

# Submit the form
submit_btn = driver.find_element(By.CSS_SELECTOR, ".captcha-submit")
submit_btn.click()

// Puppeteer
const cells = await page.$$('.captcha-grid .cell');
for (const idx of solution) {
  await cells[idx - 1].click();
}
await page.click('.captcha-submit');

Complete workflow

def solve_bls_captcha(driver, api_key):
    """Extract, solve, and submit a BLS CAPTCHA."""
    import base64, requests, time, json, re

    # 1. Extract images
    grid_cells = driver.find_elements(By.CSS_SELECTOR, ".captcha-grid img")
    images = []
    for cell in grid_cells:
        src = cell.get_attribute("src")
        if src.startswith("data:image"):
            images.append(src)
        else:
            img_data = requests.get(src).content
            b64 = base64.b64encode(img_data).decode()
            images.append(f"data:image/png;base64,{b64}")

    # 2. Extract instruction
    el = driver.find_element(By.CSS_SELECTOR, ".captcha-instruction")
    match = re.search(r'(\d{3,})', el.text)
    instruction = match.group(1)

    # 3. Submit to CaptchaAI
    data = {"key": api_key, "method": "bls", "instructions": instruction, "json": "1"}
    files = {f"image_base64_{i+1}": (None, img) for i, img in enumerate(images)}
    resp = requests.post("https://ocr.captchaai.com/in.php", data=data, files=files).json()
    task_id = resp["request"]

    # 4. Poll
    for _ in range(20):
        time.sleep(5)
        result = requests.get("https://ocr.captchaai.com/res.php", params={
            "key": api_key, "action": "get", "id": task_id, "json": "1"
        }).json()
        if result["status"] == 1:
            solution = json.loads(result["request"])
            break

    # 5. Click cells
    clickable = driver.find_elements(By.CSS_SELECTOR, ".captcha-grid .cell")
    for idx in solution:
        clickable[idx - 1].click()

    return solution

Troubleshooting

Problem	Cause	Fix
`ERROR_BAD_PARAMETERS`	Missing images or no instruction	Ensure all 9 images are valid base64 data URIs
Wrong cells selected	Incorrect cell-to-index mapping	Verify cells are numbered 1-9 left-to-right, top-to-bottom
Images not loading	Cross-origin restriction	Download images server-side and convert to base64
Empty instruction code	Instruction hidden in image	Extract instruction text or OCR the instruction image

FAQ

What does the BLS instruction code mean?

The instruction code (e.g., "664") tells the CAPTCHA which cells contain matching content. CaptchaAI analyzes it — you don't need to interpret the code yourself.