Tutorial 1: Extracting Online Reviews (e.g., Trustpilot)

​

Goal

​

Key Blocks Used

• Open Websites
• Paginate a list
• Extract Data

​

What You’ll Learn

• How to provide a list of starting URLs.
• How to configure pagination to navigate through multiple pages of reviews.
• How to define a data schema to extract structured information for each review.
• How to handle lists of items spread across multiple pages.

​

Steps

​

1. Create a New Workflow

• Navigate to your Jsonify Dashboard.
• Click the button sequence to create a new workflow: Create an empty workflow ➙ Extract.
  • This will already include Open Websites and Extract Data. We will add the Paginate a list block manually.

​

2. Configure Open Websites Block

• Select the Open Websites block.
• We’ll add a few Trustpilot company review pages as examples. In the URL input field, add the following URLs one by one (or use “Add a list of pages” from the dropdown next to the + Add button):
  • https://www.trustpilot.com/review/asana.com
  • https://www.trustpilot.com/review/monday.com
  • https://www.trustpilot.com/review/clickup.com
• Ensure the toggle switch next to each URL is ON.

Screenshot: Open Websites block configured with Trustpilot URLs

​

3. Add and Configure Paginate a list Block

• After the Open Websites block, we need to tell the agent to go through multiple pages of reviews for each company.
• To add a new block, look at the block categories at the top of the editor.
• Select the Transform category. From the dropdown list of blocks in this category, find and select Paginate a list.
• Click the + icon below the Open Websites block to add this block.
• Select the Paginate a list block to configure it.
  • How many pages to navigate through?: Set this to a reasonable number, for example, 3. This means for each company, the agent will try to load and process up to 3 pages of reviews (the initial page plus two “Next” page actions). This block handles clicking “Next page” buttons or similar pagination controls.

Paginate a list block configured for 3 pages

​

4. Configure Extract Data Block

• This block will run for each page loaded by the Open Websites and processed by the Paginate a list blocks.
• Select the Extract Data block (it should be after Paginate a list).
• What information should we scrape from each URL?: Select A list of items, as each page will contain multiple reviews.
• Define the attributes (fields) for each review:

  NAME	EXAMPLE VALUE OR A LONGER DESCRIPTION
  reviewer_name	The name of the person who wrote the review, e.g., Jane D.
  review_date	The date the review was published or experienced, e.g., Jan 15, 2024
  rating_stars	The star rating given by the user, as a number (1-5), e.g., 5
  review_title	The title or headline of the review.
  review_text	The full body text of the user's review.
  company_name	The name of the company being reviewed (extract from page or use variable if available from input).

• Add additional instructions and guidance…:
  • Extract all user reviews visible on the current page. If a field is not present for a review, leave it empty. Ensure review_text captures the complete review.

Screenshot: Extract Data block configured for extracting review details

​

5. Run Your Workflow

• Click the Run manually button (at the top right).
• The agent will now:
  1. Open the first company’s Trustpilot page.
  2. The Paginate a list block will take over, loading up to 3 pages of reviews for that company.
  3. For each of these pages, the Extract Data block will scrape all the reviews according to your schema.
  4. This process will repeat for the second company, and then the third.

​

6. Check the Results

• As the workflow runs, you’ll be redirected to the current run’s page, where results will appear.
• Once completed, click on “View as sheet”.
• You should see a table containing all the extracted reviews from the specified pages for all three companies. Each row will represent a single review with the fields you defined.