Why locally it works fine and in the server im being detected by hcaptcha?

[juanfrilla]

I'm using asynchronous seleniumbase as an API, with docker and the same proxies locally and in the server:
using these scripts:

from seleniumbase.undetected import cdp_driver


def start_xvfb_display():
    global XVFB_DISPLAY
    if XVFB_DISPLAY is None:
        from xvfbwrapper import Xvfb

        XVFB_DISPLAY = Xvfb()
        XVFB_DISPLAY.start()
        logging.info("VIRTUAL SCREEN STARTED")

async def get_webdriver_nd(proxy_data=None) -> cdp_driver.browser.Browser:
    logging.debug("Launching web browser with nodriver...")
    proxy = None
    if type(proxy_data) is dict and proxy_data != {}:
        server = proxy_data.get("server")
        username = proxy_data.get("username")
        password = proxy_data.get("password")
        if server and username and password:
            proxy = f"{username}:{password}@{server}"
        elif server:
            proxy = server
     start_xvfb_display()
    return await cdp_driver.cdp_util.start_async(
        proxy=proxy,
        browser_args=["--no-sandbox"]
    )

Like here:
#3833
It does not break anymore but the thing is that locally from Windows Subsystem for Linux (WSL) its working fine with docker + proxies +seleniumbase and on the aws EC2 server with docker + proxies + seleniumbase (same conditions) im being detected by hcaptcha, I mean, hcaptcha has a command to generate a token and inject it into the html, when I evaluate it locally (obviously automating it) it gets me a valid token and works extremely fine but on the server it appears me the captcha because Im detected as a robot and dont calculate me the token.
I think Im closer
Any recommendation?

[mdmintz]

If you're saying you can bypass hCaptcha locally with SeleniumBase via Docker, then I would be curious to see your Dockerfile and the script you used. As far as I know, Docker leaves a detectable fingerprint, making stealth automation from there unlikely. I was also under the impression that hCaptcha can't be bypassed, and therefore hCaptcha-bypass is unsupported.

Once you send over that info, I can improve capabilities, such as when running SeleniumBase from Linux servers.

Currently, there are working stealth abilities from GitHub Actions Linux servers. Eg:

• https://github.com/mdmintz/undetected-testing/actions/runs/15704681679/job/44247602812
• https://github.com/mdmintz/undetected-testing/actions/runs/15484572153/job/43596431450

[juanfrilla]

This is my dockerfile:

FROM python:3.12-slim-bookworm

# Set working directory
WORKDIR /app

# Install required system dependencies
# dumb-init kill chrome zombie processes
# xvfb for headful chrome
# gnupg verify Chrome’s signing key
RUN apt-get update && \
    apt-get install -y --no-install-recommends \
        wget \
        gnupg \
        xvfb \
        dumb-init && \
    rm -rf /var/lib/apt/lists/*

# Add Google Chrome repository and install Chrome
RUN wget -q -O - https://dl.google.com/linux/linux_signing_key.pub | apt-key add - && \
    echo "deb [arch=amd64] http://dl.google.com/linux/chrome/deb/ stable main" > /etc/apt/sources.list.d/google-chrome.list && \
    apt-get update && \
    apt-get install -y --no-install-recommends google-chrome-stable && \
    rm -rf /var/lib/apt/lists/*

# Create user and set permissions
RUN useradd --home-dir /app --shell /bin/sh chromeuser && \
    chown -R chromeuser:chromeuser /app

USER chromeuser

# Create Chrome config path to avoid warnings
RUN mkdir -p "/app/.config/chrome/Crash Reports/pending"

# Copy Python requirements and install
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy application code
COPY src ./

# Ports (adjust if necessary)
EXPOSE 8191

# Default entrypoint and command
ENTRYPOINT ["/usr/bin/dumb-init", "--"]
CMD ["python", "-u", "/app/main.py"]

[mdmintz]

That didn't work for me from a cloned SeleniumBase folder.