Running Automated Tests on CircleCI and Bitbucket Pipelines

Automated tests only deliver value when they run on every change. Here's how I set up Playwright suites on the two CI platforms I use most.

CircleCI

CircleCI's strengths are caching and parallelism. Cache the browser binaries and node_modules, then split specs across containers with the built-in test splitting.

version: 2.1
jobs:
  e2e:
    docker:
      - image: mcr.microsoft.com/playwright:v1.48.0-jammy
    parallelism: 4
    steps:
      - checkout
      - restore_cache:
          keys:
            - deps-{{ checksum "pnpm-lock.yaml" }}
      - run: pnpm install --frozen-lockfile
      - save_cache:
          key: deps-{{ checksum "pnpm-lock.yaml" }}
          paths:
            - node_modules
      - run: pnpm playwright test --shard=$((CIRCLE_NODE_INDEX+1))/$CIRCLE_NODE_TOTAL

Bitbucket Pipelines

Bitbucket Pipelines is simpler and tightly integrated with Bitbucket repos. Use the official Playwright image and Bitbucket's caches to avoid reinstalling browsers each run.

pipelines:
  pull-requests:
    "**":
      - step:
          name: E2E
          image: mcr.microsoft.com/playwright:v1.48.0-jammy
          caches:
            - node
          script:
            - pnpm install --frozen-lockfile
            - pnpm playwright test

Keep feedback fast

• Shard long suites across containers so wall-clock time stays low.
• Cache dependencies and browser binaries — installs dominate cold runs.
• Fail fast on PRs, but keep a fuller nightly run for cross-browser coverage.
• Upload artifacts (traces, videos, HTML reports) so failures are debuggable without re-running locally.

The platform matters less than the habit: every push should give the team a clear green or red within a few minutes.