Lesson 5: Add a product to the cart with Browser

Lessons 1 to 4 read pages. Some jobs need you to act on a page: pick options, click buttons, and follow the result to the next page. For that you drive a real browser. In this lesson you add a hoodie to the demo store’s cart and check what the cart says. What you’ll build: add_to_cart.py, a Playwright script that runs in Zenrows Browser, the browser for agents. It opens a product, chooses a size and a color, adds the product to the cart, reads the cart total and saves a screenshot. What you’ll learn:

How to connect Playwright to Browser with one WebSocket URL.
How to select options, click and wait for the page to respond.
How to keep a browser session short and always close it.

Time: about 10 minutes.

When to use Browser

Fetch returns a page; Browser gives you a live, cloud-hosted Chrome that you control step by step over the Chrome DevTools Protocol (CDP). For when each one fits, see Choose the right Zenrows primitive.

Install Playwright

You use Playwright’s own Python package. Browser runs Chrome in the cloud, so there’s no local browser to download:

pip install playwright

Write the cart script

Create add_to_cart.py:

Python (Playwright)

from playwright.sync_api import sync_playwright

ZENROWS_API_KEY = "YOUR_ZENROWS_API_KEY"
CONNECTION_URL = f"wss://browser.zenrows.com?apikey={ZENROWS_API_KEY}"
PRODUCT_URL = "https://www.scrapingcourse.com/ecommerce/product/abominable-hoodie/"
CART_URL = "https://www.scrapingcourse.com/ecommerce/cart/"

with sync_playwright() as p:
    browser = p.chromium.connect_over_cdp(CONNECTION_URL)
    try:
        context = browser.contexts[0] if browser.contexts else browser.new_context()
        page = context.new_page()
        page.set_default_timeout(60_000)

        page.goto(PRODUCT_URL, wait_until="domcontentloaded")
        print("Opened:", page.locator("h1.product_title").inner_text())

        page.select_option("select#size", "M")
        page.select_option("select#color", "Blue")
        page.click("button.single_add_to_cart_button")
        message = page.locator(".woocommerce-message")
        message.wait_for()
        print("Store says:", " ".join(message.inner_text().split()))

        page.goto(CART_URL, wait_until="domcontentloaded")
        total = page.locator(".wc-block-components-totals-footer-item .wc-block-components-totals-item__value")
        total.wait_for()
        item = page.locator(".wc-block-cart-item__product .wc-block-components-product-name").first
        print("In the cart:", item.inner_text())
        print("Cart total:", total.inner_text())

        page.screenshot(path="cart.png", full_page=True)
        print("Saved cart.png")
    finally:
        browser.close()

Here is what each part does:

Connect

connect_over_cdp() opens a Browser session at wss://browser.zenrows.com with your API key. From here on, page is a normal Playwright page running in the cloud.

Open the product

wait_until="domcontentloaded" continues as soon as the page’s HTML is ready, without waiting for every analytics script. The 60-second default timeout gives each step room on slower pages.

Choose options and add to cart

select_option() picks the size and the color, the same select#size and select#color dropdowns you read with CSS selectors in Lesson 2. The click submits the form, and message.wait_for() waits for the store’s confirmation.

Read the cart

The cart page renders its contents with JavaScript, so the script waits for the total to appear before it reads it. The cart remembers the hoodie because the session keeps the store’s cookies.

Close the session

browser.close() sits in a finally block, so the session ends even when a step fails. Closing the session ends its billing.

Run it

python add_to_cart.py

The whole session takes about 30 seconds:

Opened: Abominable Hoodie
Store says: View cart “Abominable Hoodie” has been added to your cart.
In the cart: Abominable Hoodie
Cart total: $69.00
Saved cart.png

Open cart.png: it shows the cart with one Abominable Hoodie, size M, color Blue, and a total of $69.00.

Make it your own

Change	How
Browse from a specific country	Add `&proxy_country=us` to the connection URL, see Country
Set how long the session lasts	Add `&session_ttl=2m` to the connection URL, see Session TTL
Use the async API or Node.js	See Playwright with Browser
Use Puppeteer instead	See Puppeteer with Browser

Ask your AI agent

With the Zenrows MCP server connected to Claude, Cursor or any MCP client, paste this prompt. Your agent picks the right Zenrows tools and settings on its own.

Use the Zenrows browser to open https://www.scrapingcourse.com/ecommerce/product/abominable-hoodie/,
choose size M and color Blue, and add it to the cart. Then open the cart, tell me the total,
take a screenshot, and close the browser session.

What you built

add_to_cart.py completes a real shopping flow in a cloud browser. You can:

Connect Playwright to Zenrows Browser with connect_over_cdp() and one URL.
Select options, click, and wait for the page to respond before you read it.
Keep sessions short and close them in a finally block.

Next lesson

Lesson 6: Run the tracker every day

Turn the Batch job into a daily schedule and get notified when each run finishes.

Start Here

Foundations Course

Fetch Recipes

Extract Recipes

Crawl Recipes

Batch Recipes

Browser Recipes

Pipelines

Use Cases

By Surface

Production

Lesson 5: Add a product to the cart with Browser

When to use Browser

Install Playwright

Write the cart script

Run it

Make it your own

Ask your AI agent

What you built

Next lesson

Lesson 6: Run the tracker every day

​When to use Browser

​Install Playwright

​Write the cart script

​Run it

​Make it your own

​Ask your AI agent

​What you built

​Next lesson

Lesson 6: Run the tracker every day

When to use Browser

Install Playwright

Write the cart script

Run it

Make it your own

Ask your AI agent

What you built

Next lesson