> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gumloop.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Web Browsing Agent

## Node Inputs

* **Starting Link**: Enter the link from which you want the web browsing agent to start the task. If not specified, the agent starts at google.com. This is optional.
* **Task**: Enter the task that you want the web browsing agent to perform. This should be a string.
* **Max. # of Actions**: The maximum number of actions that you want the web browsing agent to perform. The default is 20. This field is optional and should be an integer.
* **Show Detailed Logs**: Show detailed logs for what the web browsing agent is doing including images of what it sees and reasoning behind its actions. Note: Any images shown in these logs may be publicly accessible. Use this option for debugging purposes only. This field is optional and should be a boolean value.

## Node Output

* **Response**: Provides the response from the web browsing agent after accomplishing the task. This is a string.

# Node Functionality

The AI Web Browsing Agent is designed to simulate human-like interaction with web pages to accomplish a variety of tasks online. Here’s what it can do:

* Navigate to a specified URL to start a task.
* Perform actions like clicking, typing in text fields, scrolling up or down, waiting for a specified period, going back to the previous page, or returning to the start page.
* Process and execute defined actions based on analyzing screenshots of web pages and elements found on them.
* Provide detailed logs including screenshots for debugging purposes (if requested).
* Capable of completing tasks by interpreting simple instructions, navigating through web pages, and returning specified information.

## When To Use

This node should be used when you need to automate tasks on the web that require interaction with web pages as if a human were performing those tasks. Examples include:

* Looking up information or prices of products on various websites.
* Filling and submitting forms online.
* Navigating through multiple pages to retrieve information.
* Automating repetitive browsing tasks without the need for manual intervention.

## Important Considerations

* This node is highly experimental and consumes significant AI credits
* For repetitive web browsing requirements try the [Browser Replay](https://docs.gumloop.com/nodes/browser_extension/browser_replay) option as well
* You can use the 'configure inputs' option to expose Task, Max no. of actions, & Temperature as dynamic inputs