Web agent
A CodeWords remote browser automation tool that can perform authentication and complex multi-step tasks
Overview
The Web Agent is your AI-powered assistant for browser automation. Instead of writing complex scripts, you can simply describe your goal in plain language, and the Web Agent figures out how to achieve it by navigating websites, clicking buttons, and collecting the information you need.
Think of it as a virtual assistant that can browse the web and complete tasks for you.
When to use the web agent
Complex research tasks where you don’t know all the steps
Example: “Find the CEO’s email on the company website — check the About page, Team page, or Contact form.”
Dynamic websites that change often (e.g. LinkedIn, news sites)
Exploratory automation that requires decision-making
Example: “Search for product reviews and summarize the top 5 complaints.”
Authenticated websites behind logins (e.g. LinkedIn profiles, private dashboards)
How it works

FAQs
When should I use the Web Agent?
Use the Web Agent for complex or dynamic web tasks, especially when:
You don’t know every step ahead of time (e.g., finding a contact on a company site).
The website changes frequently or loads content dynamically.
You need to access logged-in or private pages (like LinkedIn or dashboards).
Can I watch the Web Agent in action?
Yes. You can view a live browser window as the Web Agent performs your task. This makes it easy to monitor progress, verify results, or debug if something doesn’t go as expected.
Last updated
Was this helpful?