The Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA) is designed to make automated access difficult. However, in modern software, it’s often more of a relic than a security solution. AI agents are designed to fully automate tasks with human-like competency. CAPTCHAs can make this next to impossible. CAPTCHAs don’t care if the automation is malicious or mission-critical. In some cases, your team’s AI agents can even get tripped up by your own company’s CAPTCHA system.
In 2025, CAPTCHAs still serve a purpose. That said, as AI agents become more prevalent and the system refuses to evolve, CAPTCHA solving becomes a requirement in many areas of software. In this article, we’ll explore some common tools used in automated CAPTCHA solving so your AI agents and data pipelines don’t break from a rogue CAPTCHA.
Developers are often roped into building custom tools to address these issues. In reality, production solutions with seamless integration have existed for years — and they’re probably right for your team. By the time you’re finished reading, you’ll be able to answer the following questions.
- Which types of CAPTCHAs are commonly used today?
- Which solutions are available?
- How can we implement a solution today?
- When do teams need an enterprise-level solution?
Types of CAPTCHAs
Over the last two decades, we’ve seen all sorts of new CAPTCHA types. Each new generation is designed to make access more tedious and difficult — often at the expense of user experience. There are a variety of other CAPTCHA types in use all over the web. These are some of the most common ones you might run into.
- reCAPTCHA v2: Check a box. Sometimes browsers receive an additional verification like “Select all images of a bus.”
- reCAPTCHA v3/Invisible CAPTCHA: Invisible CAPTCHAs that use JavaScript to “verify” browser legitimacy. When failed, these often result in a reCAPTCHA v2 on the user end.
- hCAPTCHA: Often more time consuming than reCAPTCHA v2. However, the same principle applies. Check the box and interact with the page.
- Text-based CAPTCHA: Text is embedded within an image. The user must enter the displayed text into an input box.
- Cloudflare turnstile: Check a box. It can sometimes result in more tedious puzzles afterward.
Most CAPTCHAs target decades-old bot behavior that doesn’t account for how bots actually behave today. Most CAPTCHA challenges are built on the following assumptions.
- Can the user read text from an image?
- Can the user interact with the browser?
- Does the user’s browser appear “normal?”
- Can the user click with precision but not too much precision?
- Can the user identify objects embedded within images?
These principles don’t accurately reflect the automation of today. With AI adoption becoming more widespread, software now has the following capabilities.
- Computer vision: Recognize and interpret objects or text from images. Models can clearly see all “images of a bus.”
- Foundational training: AI models are trained on petabytes of public web data. When it hits production, AI knows common browser setups better than most humans.
- Multistep reasoning: Break complex tasks into smaller ones for effective problem solving. When a CAPTCHA requires a checkbox and a puzzle, the agents can understand the requirements of the task.
- External tool usage: Large Language Models (LLMs) are now capable of controlling external tools. Model Context Protocol (MCP) makes it easy for AI agents to check boxes and click the correct images.
CAPTCHAs were built to block the hardcoded bots of previous generations. AI agents aren’t just tripped up by CAPTCHAs, they’re also well-equipped to solve them. More often than not, today’s CAPTCHAs are better at blocking humans than AI models. This is exactly why modern CAPTCHA handling is so reliable and efficient.
Handling CAPTCHAs with Capsolver
Now, we’ll get some hands-on experience using an automated CAPTCHA tool. The CapSolver extension runs inside your browser. It then automatically finds and solves CAPTCHAs when they’re rendered on your screen. Tools like this are useful for automated web access as well as everyday users who simply hate CAPTCHAs.
To get started, head on over to CapSolver and create an account. Once you’re finished, you should see a welcome prompt similar to the one below.
Next, you need to add credits to your account. They support a variety of payment methods such as credit/debit card, cryptocurrency and payment services such as PayPal and Alipay. You can view their pricing scale here. Just $5 is enough to solve several thousand CAPTCHAs.
Once your account is funded, it’s time to install their browser extension. It’s available on both the Chrome Web Store and the Firefox Add-ons page.
Now, feel free to open up the extension and look around. There’s an input box for your API key and a list of CAPTCHA types. Next to each CAPTCHA type, you can toggle the solver on or off using the checkbox.
Go back to the CapSolver dashboard and copy your API key. Then paste it into the extension. Once this is done, you’re ready to go.
Now we’ll visit a CAPTCHA demo page just to test out the product. As you can see in the GIF below, the CAPTCHAs on the page are solved almost instantly.
Tools like CapSolver are great for people working with local browser installations. They offer tutorials as well. Here’s one for Playwright, one of the most popular headless browsers in web development. However, when dealing with enterprise setups and cloud browsers installing extensions and add-ons isn’t feasible.
Using Web Unlockers for automated collection
Let’s take a look at a couple of enterprise solutions for CAPTCHA solving. There are all sorts of providers out there. Our examples here use Bright Data but the process remains more or less the same across the board. We’ll look at an example using Python Requests and another using a Cloud Browser. With enterprise solutions, you’re not confined to the browser.
Python Requests
The code below is very simple. We send a POST request to the network. In its body, we put our basic parameters such as URL and format. The zone parameter is somewhat unique to Bright Data. We send our API key as an authorization header. Once we’ve received our response, we write the site to an HTML file.
import requests
API_KEY = "<your-bright-data-api-key>"
endpoint = "https://api.brightdata.com/request"
payload = {
"zone": "<your-zone-name>",
"url": "https://nopecha.com/captcha/recaptcha",
"format": "raw",
}
headers = {
"Content-Type": "application/json",
"Authorization": f"Bearer {API_KEY}"
}
resp = requests.post(endpoint, json=payload, headers=headers)
with open("output.html", "w") as file:
file.write(resp.text)
After running the code, we opened up the saved HTML page inside our browser. As you can see, there are no CAPTCHAs on the page. All CAPTCHAs are solved — no browser required.
Playwright
In this next example, we use the Bright Data Python SDK to run the same basic operation using a Cloud Browser. In this case, you need your username, password, API key and zone name. We write a scrape function that connects to a website and takes a screenshot.
from brightdata import bdclient
from playwright.sync_api import Playwright, sync_playwright
client = bdclient(
api_token="<your-bright-data-api-key>",
browser_username="brd-customer-<your-username>-zone-<your-zone-name>",
browser_password="<your-password>"
)
def scrape(playwright: Playwright, url='https://nopecha.com/captcha/recaptcha'):
browser = playwright.chromium.connect_over_cdp(client.connect_browser())
try:
print(f'Connected! Navigating to {url}...')
page = browser.new_page()
page.goto(url, timeout=2*60_000)
print('Navigated! Scraping page content...')
data = page.screenshot(path="solved-captcha.png")
print(f'Scraped! Data: {data}')
finally:
browser.close()
def main():
with sync_playwright() as playwright:
scrape(playwright)
if __name__ == '__main__':
main()
We’ve got rendered text on the screen but the CAPTCHAs are gone.
Enterprise vendors
Enterprise tools make it possible to deal with CAPTCHAs at scale. When your system powers AI agents, data pipelines and other mission-critical software, solutions like these often become non-negotiable. You can’t run a production system from your local machine.
In the sections below, we’ll take a look at some of these enterprise solutions. The prices we quote are for entry-level and top tier plans. With each provider, teams receive lower prices as their usage scales.
Bright Data
Our earlier examples came from Bright Data. Both their Unlocker API and Browser API come with CAPTCHA solving and automated proxy rotation. Unlocker API pricing ranges at $1.50 for 1,000 requests ($0.0015/request) to $1,999/month for 2,000,000 at the top tier ($0.001/request). Their Browser API starts at $8/GB and goes all the way up to $1,999/month for 399GB ($5/GB).
Zyte
Zyte offers an API with optional browser rendering. They use a dynamic pricing model, so usage costs may be difficult to calculate. For unrendered requests, their pricing starts at $0.13 for 1,000 requests ($0.00013/request) on the lowest plan and goes all the way to $500/month for large-scale plans.
ZenRows
ZenRows offers a Universal Scraper API with CAPTCHA solving as well as a remote browsing product. Teams can access both products through the ZenRows monthly subscription model. Low usage plans cost $69/month and their highest tier costs $299/month.
Conclusion
For regular users, CAPTCHAs are annoying at best. When dealing with mission-critical software, they can break an entire pipeline or render your AI agents useless. With browser extensions like CapSolver, you can remove headache from daily life. For systems at scale, enterprise-grade solutions provide a stable way to keep your pipeline, your AI agents and your sanity intact.
CAPTCHAs cause headaches. With modern AI tooling, you can remove these headaches from both your personal machine and your web data infrastructure. trading agents, AI agents with autonomous search capabilities can unlock the power of real-time web data for your agentic AI system.who needs help, n8n workflows can help you turn your idea into something tangible — quickly.