NOTE: While Robot Studio is our recommended method for building robots, the Chrome extension provides an alternative for training robots directly on websites you visit.
Watch & learn
Full guide
What you'll need
A Browse AI account
Google Chrome browser
Browse AI Chrome extension (we'll help you install it if you don't have it yet)
Step 1: Start building your robot
Log into your Browse AI dashboard.
Click the Build New Robot button in the top right corner.
Choose between two options:
Extract structured data: best for one-time extractions or infrequent use.
Monitor site changes: ideal for regular tracking (daily/hourly) of website updates.
If you're not sure which option to choose we recommend selecting Extract Structured Data. You can always add monitoring to any robot as a later step.
Step 2: Enter your target website
Enter the complete URL of the webpage you want to extract data from (we call this your Original URL).
Click Start Training Robot.
Select Use Chrome Extension.
For example, we'll use Product Hunt's homepage to demonstrate extracting the top products.
Step 3: Install the Chrome extension (first-time users)
If this is your first time creating a robot, you'll be prompted to install the Browse AI Chrome extension:
Follow the Chrome prompts to add the extension.
Update your Chrome security settings following the add-on prompts as needed.
The training mode will automatically begin once installation is complete.
If this is the first time you are training a robot, you will be prompted to install the Browse AI Extension when you click on Use Chrome Extension.
Check out this brief video to see the entire process:
Step 4: Choose your capture method
To begin capturing data:
Click on the robot icon that appears on the page.
Choose one of three options:
Capture List: extract repeating items (products, listings, etc.).
Capture Text: extract specific text elements from the page.
Capture Screenshot: save an image of the page.
For our Product Hunt example, we'll select Capture List to extract multiple products.
Remember: The robot will repeat every action that you take during the training mode, so it is best to only perform scrolls and clicks when they are necessary.
Step 5: Select and name your data points
After selecting Capture List you can now select the list you want to extract:
Move your mouse over the page and notice how the robot highlights different sections.
When you see the correct list items highlighted (each product separated clearly), click to select it.
Click on each specific data point you want to extract (title, description, upvotes, etc.).
Press Enter after selecting all needed data points.
Name each field (e.g., "Product Name", "Description", "Upvote Count").
Tip: You can use the Esc key to undo any changes to the data point names.
Step 6: Configure your list settings
After naming your fields, you'll see a preview table of the captured data. Now:
Give your list a descriptive name (e.g., "Product Hunt Top Products").
Set the maximum number of items to extract.
Select the appropriate pagination type:
Click next button: for websites with next page buttons.
Scroll down to load more: for infinite scroll pages like Product Hunt.
No pagination: for single-page extractions.
Step 7: Finish and save your robot
Review your captured data in the preview.
Click Capture List to confirm.
Click on the robot icon and select Finish Recording.
Step 8: Review and approve your robot
Name your robot (e.g., "Product Hunt Top Products Tracker").
Click Save.
After saving, you'll see your extracted data.
Verify that it looks correct and click Yes, looks good.
What else can you do?
Your new robot opens up several possibilities:
Download data: go to the History tab, find your task, and click Download CSV.
Run fresh extractions: visit the Run Task tab to extract updated data anytime.
Set up a Bulk Run: automatically extract data from up to 50,000 websites with the same structure.
Set up monitoring: create a monitor to run your robot on a schedule and alert you to changes
Integrate or create automations with your data: use our integrations to send data to your favorite tools and apps.
Deep scrape: build a second robot to extract detailed information from each item in your list.