BrowserAgent: a tool for creating and running AI workflows in the browser

General Introduction

BrowserAgent is a tool that creates and runs AI workflows directly in the browser. It's easy to use and requires no code to be written; users simply describe the desired workflow and the AI is automatically generated. Its core feature is that it is completely private, all data is processed in your browser without worrying about privacy leaks. Compared to other AI tools, BrowserAgent does not charge per use, but offers an affordable model with unlimited runs. The website was built by a group of developers who wanted to break the limitations of expensive and complex AI tools and make AI easy to use for more people. it is currently receiving a lot of attention with its official debut on ProductHunt on March 15, 2025.

Function List

• Creating AI Workflows: Users can describe tasks in text and AI automatically generates runnable workflows.
• The browser runs: All operations are done locally in the browser without server support.
• No code editor: Provides a visual interface to design complex processes without knowing programming.
• Privacy: The data is not uploaded to the cloud and is processed on the user's device throughout.
• run indefinitely: One payment allows unlimited use at no extra charge.
• Diversified mission support: Suitable for form filling, data capture, price comparison and other scenarios.

Using Help

BrowserAgent is very easy to use, you don't need to install complicated software, just open your browser and you are ready to go. Below is a detailed description of how to get started and how its main features work.

How to get started

1. Access to the website
   Open your browser and go to the homepage. You will see a simple interface with a "Get Started" or "Trial" button.
2. Register or Login
   After clicking "Get Started", you will be prompted to register for an account. Just use your email and password to register. If you are a new user, you can choose to try the free trial, or just buy the affordable package (currently there is a 50% discount, the coupon code can be found on the home page).
3. Entering the workspace
   After logging in, you will be taken to a workspace. This is where you create and manage AI workflows.

Create your first AI workflow

1. Click on "New Workflow"
   In the workspace, click the "New Workflow" button in the upper left corner.
2. Description of the task
   An input box pops up, in which you are asked to describe the task in text. For example, you could type "open a website and grab the headline text" or "fill out a login form and submit it automatically". Try to be as specific as possible.
3. Generating workflows
   Click the "Generate" button and AI will create a workflow based on your description. Once generated, you will see a flowchart showing each step.
4. Adjustment process (optional)
   If you need to make changes, drag the modules in the flowchart with your mouse or click on them to adjust the details. The interface is intuitive and as simple as building blocks.
5. Running a workflow
   Click the "Run" button and BrowserAgent will execute the task in your browser. You can see the real-time progress, such as the data captured or the status of form filling.

Featured Functions Operation Procedure

data crawl

• take: Trying to grab the latest headlines from news sites.
• manipulate::
  1. Create a new workflow and enter "Open https://news.ycombinator.com and grab the header text".
  2. After the AI generation process, check for the "Navigate to URL" and "Crawl to Element" steps.
  3. In the Grab Elements module, enter the CSS selector ".titleline a" (this is the selector for the Hacker News title, if you don't know CSS, you can ask the AI to recommend it).
  4. Click "Run" and the result will be displayed on the screen, such as "so-and-so news headline".
• draw attention to sth.: If the crawl fails, adjust the selector or ask for help in the community.

Automated Form Filling

• take: Automatically log in to a site.
• manipulate::
  1. Enter the task "Open example.com/login, fill in the username 'myUser', the password 'myPass' and click on the submit button".
  2. After the AI generates the flow, check the steps: navigate to the URL, fill out the form, click the button.
  3. In the "Fill Form" module, confirm the IDs of the username and password fields (e.g. "#username" and "#password"). The "Fill in Forms" module.
  4. In the "Click Button" module, enter the ID of the submit button (e.g. "#submitBtn").
  5. After running, the browser will automatically complete the login.
• draw attention to sth.: Ensure that the site allows automation to avoid triggering CAPTCHAs.

price comparison

• take: Compare prices from multiple e-commerce sites.
• manipulate::
  1. Type "open site1.com, site2.com, grab the price of the product and organize it into a list".
  2. After the AI generation process, check the price capture module for each website.
  3. Set the CSS selector for price (e.g. ".price") for each site.
  4. Once you run it, the results will show a list of prices for you to compare.
• draw attention to sth.: You can use the "Save Results" function to export the data to a table.

practical skill

• adjust components during testing: If something goes wrong, click on "Log" to see what the problem is.
• Community Support: The site has a Discord link to join and ask questions or share workflows.
• (computer) shortcut key: Save time by pressing "Ctrl+R" to quickly run the current workflow.

caveat

• Make sure your browser supports WebAssembly, as it is used by the BrowserAgent to run the AI model.
• If the task is complex, the description should be as detailed as possible, such as specific URLs and element names.
• The data is all processed locally, but the device performance has to keep up when running big tasks.

With BrowserAgent, you'll find that it lowers the bar for AI. Whether it's grabbing data or filling out a form, if you can type, you can let AI do the work for you.