browserstack-automate

Overview

BrowserStack Automate is a hosted Selenium / Playwright / Cypress grid that exposes 3000+ real device + browser combinations (iOS, Android, Windows, macOS) via a standard WebDriver-compatible endpoint. Per browserstack.com/docs/automate/selenium.

This skill wraps BrowserStack for Selenium-style invocation; Playwright + Cypress integrations follow a different (but similar) pattern documented separately by BrowserStack.

Composes with browser-matrix-strategy-reference for matrix planning.

When to use

• Cross-browser regression across more breadth than Playwright's bundled engines support (real Safari iOS, IE11 legacy, niche Android devices, etc.).
• Tier-1 browser matrix coverage backed by SLA.
• Local-network applications testable from the cloud grid via BrowserStackLocal.

For bundled-engine matrix (Chromium / Firefox / WebKit on the runner machine), use browser-matrix-runner. For the orchestration agent that routes between local + cloud, use selenium-grid-orchestrator.

Authoring

Authentication

Per BrowserStack Automate docs, set env vars:

export BROWSERSTACK_USERNAME="your-username"
export BROWSERSTACK_ACCESS_KEY="<access-key-from-account-settings>"

Hub URL

https://hub-cloud.browserstack.com/wd/hub

Connect any WebDriver client (Selenium, WebdriverIO, Nightwatch) to this URL with the standard RemoteWebDriver-style construction.

Capabilities (W3C)

Per BrowserStack docs:

{
  "browserName": "Chrome",
  "browserVersion": "latest",
  "os": "Windows",
  "osVersion": "11",
  "bstack:options": {
    "projectName": "My App",
    "buildName": "PR-1234",
    "sessionName": "Login flow on Chrome Windows",
    "local": "false",
    "debug": "true",
    "networkLogs": "true",
    "consoleLogs": "errors"
  }
}

Standard W3C fields: browserName, browserVersion, platformName (or BrowserStack's os + osVersion non-standard).

bstack:options carries BrowserStack-specific settings:

Python example (Selenium)

import os
from selenium import webdriver

caps = {
    "browserName": "Safari",
    "browserVersion": "17",
    "os": "OS X",
    "osVersion": "Sonoma",
    "bstack:options": {
        "projectName": "my-app",
        "buildName": os.environ.get("BUILD_TAG", "local-run"),
        "sessionName": "Checkout flow on Safari macOS",
        "local": "false",
    },
}

driver = webdriver.Remote(
    command_executor=(
        f"https://{os.environ['BROWSERSTACK_USERNAME']}:"
        f"{os.environ['BROWSERSTACK_ACCESS_KEY']}"
        f"@hub-cloud.browserstack.com/wd/hub"
    ),
    options=webdriver.SafariOptions(),  # base options
)
# inject capabilities
for k, v in caps.items():
    driver.capabilities[k] = v

driver.get("https://example.com")
# ... test ...
driver.quit()

(Modern WebDriver clients prefer constructing through options + a capabilities dict; consult the chosen client's docs.)

Running

Reporting session status back

After a test, mark the session pass / fail via BrowserStack's REST API or via JS-executor:

driver.execute_script(
    'browserstack_executor: {"action": "setSessionStatus", '
    '"arguments": {"status":"passed","reason":"Login redirected as expected"}}'
)

Or "status":"failed","reason":"..." on failure.

This drives the BrowserStack dashboard's pass / fail metrics + filtering.

Local testing (BrowserStackLocal)

To run tests against localhost / internal environments:

# Download the BrowserStackLocal binary from browserstack.com
./BrowserStackLocal --key "$BROWSERSTACK_ACCESS_KEY" --daemon start
# Sessions with bstack:options.local = "true" now tunnel
./BrowserStackLocal --key "$BROWSERSTACK_ACCESS_KEY" --daemon stop

Or via Docker:

docker run --name bstacklocal -d --rm \
  browserstack/local --key "$BROWSERSTACK_ACCESS_KEY"

Parallel session limits

BrowserStack plans limit parallel sessions (typically 5-50 per plan). Per their docs, queue overflow blocks subsequent sessions until earlier ones complete. Plan accordingly:

from concurrent.futures import ThreadPoolExecutor

MAX_PARALLEL = 5  # match plan

with ThreadPoolExecutor(max_workers=MAX_PARALLEL) as exe:
    for case in cases:
        exe.submit(run_case, case)

Parsing results

BrowserStack session reports include:

• Session video (replay any failure)
• Network HAR (if networkLogs: true)
• Browser console logs (if consoleLogs: errors|warnings|info|verbose)
• Selenium logs
• Visual debugging (screenshots at each command)

Per the BrowserStack docs, retrieve sessions via REST API:

curl -u "$BROWSERSTACK_USERNAME:$BROWSERSTACK_ACCESS_KEY" \
  "https://api.browserstack.com/automate/sessions/<session-id>.json"

Feed failure videos + HAR to bug-report-from-failure for triage.

CI integration

# .github/workflows/cross-browser.yml
on: pull_request
jobs:
  bstack:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        browser:
          - { name: Chrome, version: latest, os: Windows, osVersion: "11" }
          - { name: Safari, version: "17", os: "OS X", osVersion: Sonoma }
          - { name: Firefox, version: latest, os: Windows, osVersion: "11" }
          - { name: Edge, version: latest, os: Windows, osVersion: "11" }
    steps:
      - uses: actions/checkout@v5
      - name: Run cross-browser tests
        env:
          BROWSERSTACK_USERNAME: ${{ secrets.BROWSERSTACK_USERNAME }}
          BROWSERSTACK_ACCESS_KEY: ${{ secrets.BROWSERSTACK_ACCESS_KEY }}
          BSTACK_BROWSER: ${{ matrix.browser.name }}
          BSTACK_VERSION: ${{ matrix.browser.version }}
          BSTACK_OS: ${{ matrix.browser.os }}
          BSTACK_OS_VERSION: ${{ matrix.browser.osVersion }}
          BUILD_TAG: pr-${{ github.event.pull_request.number }}
        run: pytest tests/e2e/ --bstack

Anti-patterns

Limitations

• Cost. Plan tiers limit parallel sessions; full-matrix runs on every PR can be expensive - tier the matrix.
• Setup latency. Real-device sessions have 5-30s startup; not optimal for short feedback loops.
• Network shape. Cloud-grid latency is higher than local; some timing-sensitive tests behave differently.
• Real-device matrix changes. BrowserStack adds + retires devices; tests pinned to specific versions may need updates.
• Internal network requires BrowserStackLocal. Adds setup + tunnel-stability concerns.

References

• BrowserStack Automate Selenium docs - browserstack.com/docs/automate/selenium.
• BrowserStack capabilities reference - browserstack.com/automate/capabilities (Cloudflare-protected; cite by stable URL).
• BrowserStackLocal - browserstack.com/local-testing/automate.
• W3C WebDriver specification - w3.org/TR/webdriver2/.
• Composes: browser-matrix-strategy-reference.
• Sibling skills: saucelabs-automate, lambdatest-automate, selenium-grid-4-runner.
• Existing orchestrator: selenium-grid-orchestrator - routes between local Selenium Grid + BrowserStack + Sauce Labs + LambdaTest.