Learning to Use Nodes - Playwright Node

The Playwright node connects to a remote Playwright server (or a browser that supports CDP) to control a browser automatically. It runs a user-defined Playwright script for browser automation.

While the Web Request node is suited for talking to servers that return structured responses like a REST API, the Playwright node is for when you need real browser actions (navigating pages, clicking, typing, scrolling, extracting text/data, etc.) on a web page that has no API available. Agentria doesn't run the browser itself — it connects to a remote Playwright server you've already set up and controls the browser via Python code.

Adding the Node

Click +Add Node on the canvas, then find and place the Playwright node under the Utility/Productivity category.

Fields

The Playwright node provides the following two fields by default.

You can also add the following option variables as needed, by clicking +Add Variable in the Node Editor and checking them.

Writing the Playwright Code

Once you enter the server address in the remote_url field, you can start writing your automation logic directly in the code field. The code uses the async Playwright API, and the built-in objects below are already injected and connected — you don't need to write any code to connect to the server yourself.

• browser — the browser instance

• context — the BrowserContext (cookies, storage, etc.)

• page — the current page (used most often)

• logger — for log output

• print — for console output

• base64 — base64 encoding/decoding

Your code must follow these two rules.

• You must return the result with a return statement — the returned value is stored in the result variable.

• Every Playwright API call must use await to call it asynchronously.

Here's the example code the node provides by default.

# Available objects: browser, context(BrowserContext), page, logger, print
# Available modules: base64

# Navigate to a webpage and capture title
await page.goto("<https: example.com="">")

# Get page information
title = await page.title()
url = page.url
storage_state = await context.storage_state()

# Return results
return {
  "result": {
    "title": title,
    "url": url,
    "storage_state": storage_state
  }
}</https

# Available objects: browser, context(BrowserContext), page, logger, print
# Available modules: base64

# Navigate to a webpage and capture title
await page.goto("<https: example.com="">")

# Get page information
title = await page.title()
url = page.url
storage_state = await context.storage_state()

# Return results
return {
  "result": {
    "title": title,
    "url": url,
    "storage_state": storage_state
  }
}</https

# Available objects: browser, context(BrowserContext), page, logger, print
# Available modules: base64

# Navigate to a webpage and capture title
await page.goto("<https: example.com="">")

# Get page information
title = await page.title()
url = page.url
storage_state = await context.storage_state()

# Return results
return {
  "result": {
    "title": title,
    "url": url,
    "storage_state": storage_state
  }
}</https

Common Usage Examples

Here are some frequently used page object calls.

• await page.goto("URL") — navigate to a page

• await page.click("selector") — click an element

• await page.fill("input", "value") — type text into an input field

• await page.screenshot() — capture a screenshot

• await page.inner_text("selector") — extract text (scraping)

• await page.wait_for_selector("selector") — wait for an element to load

Testing and Checking the Result

Click RUN TEST to test the node and confirm it works correctly. The return value is stored in the result variable and passed on to the next node.