Robot Framework & Playwright
Browser Library allows you to run Playwright tests with Robot Framework. By using TestingBot, you will be able to run Robot tests using Playwright on TestingBot's browser grid. The advantage is that you do not have to set up and maintain a fleet of different browsers, versions and operating systems. Another advantage is that you can run multiple tests, in parallel, on TestingBot's grid.
Example
Below is an example on how to connect to the TestingBot browser grid, using Browser Library's Connect To Browser. The connection is made from a small Python library, so your TestingBot key and secret never end up in Robot Framework's output.xml or log.html (see Keep your credentials out of the logs).
pip install robotframework robotframework-browser
rfbrowser init chromium
Save the code below in a TestingBot.py file:
import json
import os
from datetime import timedelta
from urllib.parse import quote
from Browser.utils.data_types import SupportedBrowsers
from robot.libraries.BuiltIn import BuiltIn
# Playwright engine that drives each TestingBot browser
ENGINES = {"chrome": "chromium", "edge": "chromium", "firefox": "firefox", "safari": "webkit"}
def connect_to_testingbot(browser, platform, version="latest", name="", build=""):
"""Connects the Browser library to a TestingBot browser.
The key and secret are read from the TB_KEY and TB_SECRET environment
variables and only exist inside Python, so they are never logged.
"""
capabilities = {
"browserName": browser.lower(),
"browserVersion": version,
"platform": platform,
"tb:options": {
"key": os.environ["TB_KEY"],
"secret": os.environ["TB_SECRET"],
"name": name,
"build": build,
},
}
endpoint = ("wss://cloud.testingbot.com/playwright?capabilities="
+ quote(json.dumps(capabilities)))
BuiltIn().get_library_instance("Browser").connect_to_browser(
endpoint, SupportedBrowsers[ENGINES[browser.lower()]], timeout=timedelta(minutes=2))
Then save the test in a file, for example test.robot:
*** Settings ***
Library Browser enable_playwright_debug=disabled
Library TestingBot.py
Test Setup Connect To TestingBot chrome WIN11 latest
... name=${TEST NAME} build=robot-playwright
Test Teardown Report Status And Close
*** Test Cases ***
Playwright Test
New Page https://playwright.dev/
Get Title matches Playwright
Get Attribute "Get started" href == /docs/intro
Click "Get started"
Get Url matches .*intro
*** Keywords ***
Report Status And Close
${passed}= Evaluate $TEST_STATUS == 'PASS'
&{arguments}= Create Dictionary passed=${passed} reason=${TEST MESSAGE}
&{command}= Create Dictionary action=setSessionStatus arguments=${arguments}
${json}= Evaluate json.dumps($command) modules=json
Evaluate JavaScript ${None} () => {} arg=testingbot_executor: ${json}
Close Browser
Run the test with your TestingBot credentials in the environment:
export TB_KEY=add-your-TestingBot-key
export TB_SECRET=add-your-TestingBot-secret
robot --outputdir results test.robot
This example will open a Chrome browser on Windows 11 using Playwright on TestingBot, navigate to the Playwright website and do some verifications. The teardown sends the result to TestingBot with the testingbot_executor: setSessionStatus command, so the test shows up as passed or failed in your dashboard.
- The client-side engine must match the remote browser:
chromiumfor Chrome and Edge,firefoxfor Firefox andwebkitfor Safari. TheENGINESlookup takes care of that. - The connect timeout is raised to two minutes, because a remote browser (Firefox in particular) can take longer than Browser Library's 30 second default to start, and a session may wait in the queue for a free slot.
Keep your credentials out of the logs
Robot Framework logs every variable assignment. If you build the capabilities in Robot Framework syntax (for example &{caps}= Create Dictionary key=... secret=...), your TestingBot secret is written to output.xml and log.html, which CI systems often publish as build artifacts. Building the endpoint in Python, as above, avoids that.
Browser Library also writes its own playwright-log.txt next to the Robot Framework output. That file logs the full endpoint URL, including the secret, so the example imports the library with enable_playwright_debug=disabled. Running the example above on TestingBot with that setting left 0 occurrences of the secret in the output directory.