14 min
Playwright MCP vs CLI vs Agents: What to Use in 2026
Playwright has three ways to talk to AI: MCP, CLI, and Test Agents. Here's the decision framework an enterprise SDET uses to pick the right one for 2026.
§ Journal · Archive
Blog
12 min
Claude Code Has 2 Primitives, Not 3 — Use Skills First
Most engineers think Claude Code has three primitives. It actually has two — skills and subagents. Here's when to use which, with token-cost benchmarks.
11 min
Playwright's New CLI Debugger Finally Works in Headless CI
Playwright 1.59 ships --debug=cli and scriptable trace analysis. The first debug surface that actually works where enterprise tests fail — headless CI.
11 min
Playwright 1.59's Screencast API Ends Useless Test Videos
Playwright 1.59 ships a Screencast API with burned-in action annotations and chapter markers. Microsoft sells it for AI agents — the real win is CI triage.
24 min
Stop Sharing Test Environments — Build One You Control
Shared staging environments are the #1 source of flaky UI tests in enterprise teams. Here's the 5-pillar blueprint for building a controlled test environment.
11 min
We Cut 150 Min of Test Setup with 3 Java Classes
How we built a Playwright API test data layer in Java — replacing UI setup that burned 150 minutes per run with 3 classes and a legacy API wrapping pattern.
13 min
Contract Testing vs API Testing — What's the Difference?
Contract testing and API testing look similar but they catch very different bugs. Here's where contract testing fits next to unit, API, and E2E tests.
13 min
Parallel Execution Without the Refactor Tax
Thread-level parallelism demands ThreadLocal and full framework rewrites. Process-level parallelism doesn't. Here's the difference and when to use each.
9 min
Stop Using locator() for Everything in Playwright
Most Playwright teams still use CSS selectors via locator() out of habit. getByRole and getByText find elements the way users do — and survive redesigns.
12 min
Shared Test Users Are Sabotaging Your Parallel Suite
How shared test accounts create race conditions in parallel Playwright runs, and the 3 isolation patterns that eliminated our 12% failure rate overnight.
8 min
Stop Launching a Browser Per Test in Playwright
Launching a new browser per test in Playwright wastes 3-6 seconds each time. BrowserContext gives you the same isolation in milliseconds. Here's how to switch.
10 min
Playwright's storageState() Is Contaminating Your Tests
storageState() saves cookies, localStorage, and sessionStorage in one blob. Here's how it silently poisons parallel tests and how to strip it to cookies only.
15 min
Selenium vs Playwright in 2026: Same Test, Two Frameworks
We build the exact same Java test in both Selenium and Playwright side by side. See the real code differences in setup, locators, network mocking, and debugging
6 min
The Browser Errors Your Test Suite Never Catches
Your UI tests pass green while the console throws errors. Learn to catch JavaScript and page errors in Selenium and Playwright Java — before users do.
12 min
Your Test Suite Is Slow for 5 Reasons — Not Just One
Parallelism alone won't save your slow test suite. Here are five layers of optimization — from test design to cloud infrastructure — ranked by real impact.
9 min
Shift-Left Is a Lie If Only QA Shifted
Shift-left promised earlier testing. In practice, it shifted QA's workload without shifting developer accountability. Here's what actually needs to change.
13 min
Shared Session Cookies Corrupted Our Parallel Tests
Eight Playwright BrowserContexts were properly isolated, but they all loaded the same session cookie — so the server treated 8 parallel threads as one.
12 min
Self-Healing Locators Aren't Magic — Here's How They Work
Most self-healing test tools use simple fallback chains, not real AI. Learn the 4 approaches to AI locator repair and which one fits your framework best.
13 min
Your Green Test Suite Is Hiding Real Bugs
Handling unexpected alerts or removing assertions to make tests pass doesn't fix anything — it buries bugs. Here's how to spot and stop these patterns.
5 min
Playwright's onceDialog Has a Silent Handler Leak
Why onceDialog stays registered when no dialog fires, silently eating the next dialog from an unrelated action — and the onDialog/offDialog pattern that fixes it.
13 min
The Flaky Test Isn't Flaky — It's a Race Condition
How a 'flaky' Playwright test exposed a shared test user race condition in our parallel suite, and the isolation patterns that fixed it for good.
12 min
Selenium's Alert Handling Crashed Our Parallel Suite
How UnhandledAlertException broke 8-thread parallel execution and why Playwright's event-driven dialog model avoids that entire failure pattern in practice.
3 min
Playwright's onceDialog Saved Me From a Handler Leak — Until It Didn't
How onceDialog fixes permanent handler leaks from onDialog — and the one case where it creates a new problem. Includes the onDialog/offDialog pattern for dialogs that might not fire.
12 min
Stop Emailing Test Reports — Host Allure on Jenkins
Replace emailed Extent Report HTML files with a persistent Allure portal on Jenkins, so the team gets one URL, full history, and zero downloads to manage.
10 min
Host a Team Report Portal with Allure Docker Service
Set up a standalone Allure report portal with Docker that any CI tool can push to, so the whole team can review results without logging into Jenkins daily.
10 min
Your Copilot Isn't Dumb — You're Starving It of Context
Enterprise GitHub Copilot stuck on an older model? Three files give it project context, reusable prompts, and path-specific rules that improve output.
11 min
Task Scheduler Runs Our Automation Server — Here's How
Five real ways I use Windows Task Scheduler to keep an enterprise Playwright automation server healthy — from nightly test runs to disk alerts and cleanup.
11 min
2 Minutes a Day Cost Me More Than You Think
I automated my enterprise VPN login with Python and saved 10+ hours a year. But the real win wasn't time — it was the focus and brain energy I never got back.
11 min
Our Enterprise Approved AI — And Why It's the Biggest Risk
Enterprises lock teams into outdated AI models for safety. The irony? Older, less capable models produce worse code and create more risk than they prevent.
5 min
From Selenium Wrappers to Playwright Locators
How Playwright's locator API eliminates the need for custom wrapper classes we spent weeks building in Selenium, and when wrappers still make sense today.
11 min
I Let AI Write My Test Suite — Here's What Broke
A hands-on experiment with AI-generated Playwright tests: where LLMs save time, where they create false confidence, and the review workflow that works.
8 min
How We Stopped Arguing About Whose Tests Failed
Use Cucumber squad tags and Extent Reports to route test failures to the right team automatically, with full implementation details for Appium projects.
7 min
Your Selectors Keep Breaking — Start Using Text
Why text-based locators reduce test maintenance by up to 60%, and how to use them in Selenium and Playwright with real enterprise UI examples that last.
5 min
Managing UI Text in Test Automation
When to use properties files, enums, or spreadsheets for UI text in test automation, and how to keep assertions maintainable across locales and releases.
8 min
Testing Webhooks Requires a Different Strategy
Why your API testing patterns fail for webhooks, and the listener-based strategy that fixed our CI pipeline after a full week of production-style debugging.
5 min
Passing Inputs to Tests with GitHub Actions
Use workflow_dispatch inputs to choose platform, suite, and release data at runtime without cloning workflows or hardcoding separate jobs for each scenario.
5 min
Send Test Results to Slack Without Losing the Signal
A practical pattern for posting color-coded test summaries to Slack from GitHub Actions, with thresholds and summaries that help teams notice failures fast.
9 min
What Makes a Good Test?
A practical guide to writing tests that catch broken behavior, fail for useful reasons, and keep earning trust in noisy enterprise CI pipelines over time.
6 min
Migrating Off Cypress? Here's When to Keep It
An honest take on Cypress vs Playwright migrations from an SDET who's done three, including the signals that tell you not to migrate your suite just yet.
6 min
The Async Trap Behind Flaky Playwright Tests
The 3 async mistakes that cause flaky Playwright tests after a Selenium migration, and how we fixed a 23% intermittent failure rate in our CI pipeline.
11 min
Thread Safety in Parallel Tests: The 3-Day Bug
How a shared WebDriver instance caused phantom failures in our 800-test parallel suite, and the ThreadLocal pattern that finally stabilized the whole run.
7 min
What the SDET Role Looks Like in the Age of AI
How AI is changing the SDET role in practice, from faster test generation and log analysis to stronger review, evaluation, and system design judgment.
4 min
Tight vs Loose Coupling in Java Test Design
A practical explanation of tight and loose coupling in Java, with examples that show why dependency boundaries matter for maintainability and testing.
8 min
XPath text() vs Dot — Why Your Text Match Fails
The real difference between XPath text(), dot, contains(), and normalize-space() for test automation — with examples that explain real flaky failures.