[Python] Is it possible to use browser-use with undetected-chromedriver?

I’m trying to drive an automated web search with the browser-use library, but I need to swap out the default Playwright for an undetected-chromedriver instance so that I can interact with pages guarded by reCAPTCHA (e.g. https://recaptcha-demo.appspot.com/recaptcha-v2-invisible.php).

undetected-chromedriver: https://github.com/ultrafunkamsterdam/undetected-chromedriver browser-use: https://github.com/browser-use/browser-use

When I replace the browser in my script with an undetected_chromedriver.Chrome(...), I see failures in the agent steps and eventually get:

ERROR [agent] ❌ Stopping due to 3 consecutive failures
None
INFO [uc] ensuring close
Exception ignored in: <function _ProactorBasePipeTransport.__del__ at 0x0000024128D00FE0>
Traceback (most recent call last):
File "C:\Program Files\Python311\Lib\asyncio\proactor_events.py", line 116, in __del__
_warn(f"unclosed transport {self!r}", ResourceWarning, source=self)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Program Files\Python311\Lib\asyncio\proactor_events.py", line 80, in __repr__
info.append(f'fd={self._sock.fileno()}')
^^^^^^^^^^^^^^^^^^^


Below is a simplified version of my script:

import os
from dotenv import load_dotenv
from langchain_google_genai import ChatGoogleGenerativeAI
from browser_use import Agent, BrowserSession, Controller
from pydantic import BaseModel
import undetected_chromedriver as uc
from selenium.webdriver.common.by import By
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC

load_dotenv()

# Initialize Gemini LLM
llm = ChatGoogleGenerativeAI(
model=os.getenv("GEMINI_MODEL_NAME"),
api_key=os.getenv("GEMINI_API_KEY"),
)

# Launch undetected Chrome
driver = uc.Chrome(headless=False)
browser_session = BrowserSession(driver=driver)

class ResponseResult(BaseModel):
response: str

controller = Controller(output_model=ResponseResult)

async def main():
try:
driver.get('https://recaptcha-demo.appspot.com/recaptcha-v2-invisible.php')
WebDriverWait(driver, 20).until(EC.presence_of_element_located((By.TAG_NAME, 'body')))
await asyncio.sleep(2)

agent = Agent(
task="Find the text input fields, enter any data, send, and extract the response",
llm=llm,
browser_session=browser_session,
initial_actions=[]
)

result = await agent.run()
print(result.final_result())

except Exception as e:
print(f"Error during execution: {e!s}")
finally:
driver.quit()
await browser_session.close()

if __name__ == "__main__":
asyncio.run(main())


Expected behavior: The agent should be able to open the invisible reCAPTCHA demo page, perform the actions, and return the extracted response text.

Actual behavior: The agent fails at Step 1 immediately, logging consecutive failures and then an unclosed transport warning from asyncio.

Environment: Python Python 3.11.1 undetected-chromedriver 3.5.5 browser-use: 0.2.5 OS: Windows 10

Continue reading...