SYSTEM: Command browse_website returned: Error

[thomasmunoz]

Duplicates

• [X] I have searched the existing issues

Steps to reproduce 🕹

appears when it tries to brows a web site

Current behavior 😯

No response

Expected behavior 🤔

No response

Your prompt 📝

# Paste your prompt here

[YHrhhh]

NEXT ACTION: COMMAND = browse_website ARGUMENTS = {'url': 'https://status.openai.com/', 'question': 'Are there any reported issues with the OpenAI API currently?'} SYSTEM: Command browse_website returned: Error: Message: unknown error: cannot find Chrome binary Stacktrace: Backtrace: GetHandleVerifier [0x0065DCE3+50899] (No symbol) [0x005EE111] (No symbol) [0x004F5588] (No symbol) [0x00510AAB] (No symbol) [0x0050F479] (No symbol) [0x00541FFE] (No symbol) [0x00541CEC] (No symbol) [0x0053B6F6] (No symbol) [0x00517708] (No symbol) [0x0051886D] GetHandleVerifier [0x008C3EAE+2566302] GetHandleVerifier [0x008F92B1+2784417] GetHandleVerifier [0x008F327C+2759788] GetHandleVerifier [0x006F5740+672048] (No symbol) [0x005F8872] (No symbol) [0x005F41C8] (No symbol) [0x005F42AB] (No symbol) [0x005E71B7] BaseThreadInitThunk [0x76FD6A39+25] RtlGetFullPathName_UEx [0x77C87C9D+173] RtlGetFullPathName_UEx [0x77C87C6B+123] XIAOMING THOUGHTS: Unfortunately, the 'browse_website' command failed to execute because Chrome is not installed. Therefore, I will manually navigate to the OpenAI GPT-3 API status page and review the information to make sure the API is operational and ready to support my article summarizer agent. REASONING: Reviewing the OpenAI GPT-3 API status page is important for the article summarizer agent to function properly. Even though the manual approach will take slightly longer, I will still have access to the information I require to make informed decisions on what to do next.

[adevinwild]

I'm on MacOS and I use Arc browser, after looking at the code of the browse_website function, it uses selenium and selenium needs the Chrome drivers apparently to work.

Source: https://www.selenium.dev/documentation/webdriver/getting_started/install_drivers/

Solution

I installed Google Chrome on my machine using brew and the problem was solved 🎉

Note : I think that it should work on all OS 👍

Tell me if it's OK for you too

[JingyuQian]

I had the problem while deploying on a barebone VM. You need to have both chromedriver and chrome binary on your system.

The former is easy since the program will install it for you if there isn't one. The latter, you need to install it yourself:

wget https://dl.google.com/linux/direct/google-chrome-stable_current_amd64.deb
sudo dpkg -i google-chrome-stable_current_amd64.deb

(Optionally you might need to fix some broken dependencies for chrome: sudo apt --fix-broken install)

NOTE if your system doesn't have X display, chrome will crash. In that case refer to #1517 to add 2 launch options in the code.

[aetherwu]

I've got empty errors：

···Command browse_website returned: Error: Message:···

[413x45h4w]

My autogpt gets stuck on an endless loop here as well.

[oldwhite83]

I've got empty errors： ···Command browse_website returned: Error: Message:···

Hello, have you solved this problem? I have encountered the same problem

Hello, have you solved this problem? I have encountered the same problem

[adevinwild]

Try to :

• Install chrome (or chrome drivers)
• Check that you have OpenAI credits( Usage tab)

[413x45h4w]

@adevinwild those aren't the issue for me

[kratoszmx]

I've got empty errors：

···Command browse_website returned: Error: Message:···

1、在workspace.py中添加3行代码

os.environ["http_proxy"] = "http://127.0.0.1:7890"
os.environ["https_proxy"] = "http://127.0.0.1:7890"
os.environ["all_proxy"] = "socks5://127.0.0.1:7890"

2、修改browse_website函数： modify browse_website function： ./autogpt/commands/web_selenium.py:

“def` browse_website(url: str, question: str) -> tuple[str, WebDriver]:
    if 'http_proxy' in os.environ:
        del os.environ['http_proxy']

    if 'https_proxy' in os.environ:
        del os.environ['https_proxy']

    if 'all_proxy' in os.environ:
        del os.environ['all_proxy']

    """Browse a website and return the answer and links to the user

    Args:
        url (str): The url of the website to browse
        question (str): The question asked by the user

    Returns:
        Tuple[str, WebDriver]: The answer and links to the user and the webdriver
    """
    driver, text = scrape_text_with_selenium(url)
    add_header(driver)
    summary_text = summary.summarize_text(url, text, question, driver)
    links = scrape_links_with_selenium(driver, url)

    # Limit links to 5
    if len(links) > 5:
        links = links[:5]
    close_browser(driver)

    os.environ["http_proxy"] = "http://127.0.0.1:7890"
    os.environ["https_proxy"] = "http://127.0.0.1:7890"
    os.environ["all_proxy"] = "socks5://127.0.0.1:7890"

    return f"Answer gathered from website: {summary_text} \n \n Links: {links}", driver

[Pwuts]

Most of these issues should be resolved by #1473. If your issue still exists in current master, please attach your log to an issue about your specific problem.

[oldwhite83]

ssl_client_socket_impl.cc(992)] handshake failed; returned -1, SSL error code 1

这些是sll证书错误，可以通过配置Chrome 的option 屏蔽到这些错误，不屏蔽也没事。

[WangPeterXF]

想问一下在workspace.py中添加的3行代码： (I would like to ask about the 3 lines of code added to workspace.py:) os.environ["http_proxy"] = "http://127.0.0.1:7890" os.environ["https_proxy"] = "http://127.0.0.1:7890" os.environ["all_proxy"] = "socks5://127.0.0.1:7890"

是添加在什么位置吗？还是说添加在任何位置都行？ (Is it added in any position? Or can I add it anywhere?)

[WangPeterXF]

finally i fixed this one.

Most of these issues should be resolved by #1473. If your issue still exists in current master, please attach your log to an issue about your specific problem.

finally i fixed this error. SYSTEM showed an error, but no error message. My autogpt running under CMD, win 10. I think this is a particular error for those who use openai through VPN.

How i fix: Step 1: add modify code in "workspace.py" from future import annotations #this line already exist

import os

os.environ["http_proxy"] = "http://127.0.0.1:7890" os.environ["https_proxy"] = "http://127.0.0.1:7890" os.environ["all_proxy"] = "socks5://127.0.0.1:7890"

Step 2: modify ./autogpt/commands/web_selenium.py:

"""Selenium web scraping module.""" from future import annotations

import os #add this line import logging . . . then def browse_website(url: str, question: str) -> tuple[str, WebDriver]: if 'http_proxy' in os.environ: #add this line del os.environ['http_proxy'] #add this line

if 'https_proxy' in os.environ: #add this line
    del os.environ['https_proxy'] #add this line

if 'all_proxy' in os.environ: #add this line
    del os.environ['all_proxy'] #add this line

hope this help other people

[kratoszmx]

finally i fixed this one.

Most of these issues should be resolved by #1473. If your issue still exists in current master, please attach your log to an issue about your specific problem.

finally i fixed this error. SYSTEM showed an error, but no error message. My autogpt running under CMD, win 10. I think this is a particular error for those who use openai through VPN.

How i fix: Step 1: add modify code in "workspace.py" from future import annotations #this line already exist

import os

os.environ["http_proxy"] = "http://127.0.0.1:7890" os.environ["https_proxy"] = "http://127.0.0.1:7890" os.environ["all_proxy"] = "socks5://127.0.0.1:7890"

Step 2: modify ./autogpt/commands/web_selenium.py:

"""Selenium web scraping module.""" from future import annotations

import os #add this line import logging . . . then def browse_website(url: str, question: str) -> tuple[str, WebDriver]: if 'http_proxy' in os.environ: #add this line del os.environ['http_proxy'] #add this line

if 'https_proxy' in os.environ: #add this line
    del os.environ['https_proxy'] #add this line

if 'all_proxy' in os.environ: #add this line
    del os.environ['all_proxy'] #add this line

hope this help other people

对的，应该import一下os库

[oppiet30]

I've got empty errors：

···Command browse_website returned: Error: Message:···

You have to be running in a gui and have chrome browser installed or set the browser option in the .env file. Or set the browser variable to Firefox or Opera.

[herrinx]

This was annoying me. I'm on windows. This fixed it for me.

IF you run into it, do this.

Go here and download the appropriate webdriver for your version of chrome. https://chromedriver.chromium.org/downloads Extract it and save the binary to your AI root folder or wherever.

Then run this command from a command prompt. setx PATH "%PATH%;C:\WebDriver\bin" where the c:\ whatever is your ai folder. For me the command was setx PATH "%PATH%;K:\AI"

reboot or restart cmd prompt

From a command prompt, type echo %PATH% ensure it includes that folder.

From a command prompt, type chromedriver.exe

If it runs, you are good to go

[gsultanshah]

brew

Thanks, it works