behave-playwright-browserstack (BrowserStack SDK + Playwright)

This repo shows how to run behave tests on BrowserStack using the BrowserStack Python SDK and Playwright Python. The SDK handles capability injection, BrowserStack routing for Playwright launches, parallelization, and BrowserStack Local for you — you describe platforms once in browserstack.yml and run the test command unchanged.

Setup

• Clone this repo
• Install dependencies (creates the BrowserStack SDK CLI on PATH and downloads Playwright Chromium)

  pip install -r requirements.txt
  playwright install chromium

• Set BROWSERSTACK_USERNAME and BROWSERSTACK_ACCESS_KEY as environment variables, or replace userName and accessKey directly in browserstack.yml with your BrowserStack Username and Access Key. Env vars take precedence.

Running your tests

There are two sample scenarios in features/:

• features/sample.feature — drives https://www.bstackdemo.com (a public site) and adds a product to the cart.
• features/local.feature — drives http://bs-local.com:45454/ through the BrowserStack Local tunnel; verifies the page title contains "BrowserStack Local".

browserstack.yml enables browserstackLocal: true, so the SDK starts and stops the BrowserStack Local tunnel for you on every run — no manual binary lifecycle.

Sample test (public site)

Runs in parallel across the 3 Playwright browser engines (chromium / firefox / webkit) declared in browserstack.yml:

browserstack-sdk behave features/sample.feature

Local test (private / localhost host)

Start a local HTTP server first — features/local-html/ contains a tiny page titled "BrowserStack Local Test Page":

python3 -m http.server 45454 --directory features/local-html

Then in a separate terminal:

browserstack-sdk behave features/local.feature

bs-local.com is a hostname BrowserStack Local resolves to your machine inside the remote browser — for your own app, point your scenarios at http://bs-local.com:<port>/ instead of a public URL.

Understand how many parallel sessions you need by using our Parallel Test Calculator.

Alternatively the variables can be set in the environment using env or your CI framework (like GitHub Actions or Jenkins). See .github/workflows/build.yml for a GitHub Actions example — it runs on workflow_dispatch (manual trigger) with a commit SHA input and posts a status check back to that commit.

How the SDK changes things

• One browserstack.yml declares platforms; the SDK picks them up automatically.
• The SDK runs platforms in parallel for you — no hand-rolled parallel runner; the SDK forks one behave run per (platform × parallelsPerPlatform) cell.
• The SDK monkeypatches Playwright's browser launches — the test code calls chromium.launch() and the SDK transparently routes the launch to the per-platform browser configured in browserstack.yml (chromium, firefox, or webkit). No chromium.connect(wss_url) plumbing is needed in customer code.
• The CLI is browserstack-sdk behave … — wraps behave and injects the SDK at runtime.

Repo layout

.
├── browserstack.yml           # SDK config: credentials, 3 parallel platforms, Local toggle, reporting
├── requirements.txt
└── features/
    ├── sample.feature         # bstackdemo add-to-cart scenario
    ├── local.feature          # BrowserStack Local tunnel scenario
    ├── local-html/
    │   └── index.html         # static page served on :45454 for local.feature
    ├── environment.py         # behave hooks: launch browser, hand to context.page
    └── steps/
        ├── sample_steps.py
        └── local_steps.py

Further Reading

• behave
• Playwright Python
• BrowserStack documentation for Playwright
• BrowserStack Python SDK on PyPI

Happy Testing!