Test-Lab now combines axe-core static analysis with an AI agent that tests keyboard navigation, focus management, and interactive patterns - in a single run.
Microsoft now ships two ways to give AI agents browser control - MCP and CLI. One streams rich snapshots into context. The other saves state to disk. Here's when to use each, with real token numbers.
A practical, evidence-driven workflow for debugging flaky AI browser tests in CI. Diagnose stale DOM, auth drift, bot protection, timing, and agent loops fast.
Datacenter proxies get flagged. We integrated NordVPN with a Docker sidecar pattern to give you real VPN IPs that geo-restricted sites actually trust.
Your app behaves differently in Germany than the US. Test that. Route AI tests through 110+ countries with proxy and VPN integration.
Test-Lab now supports mobile device emulation. Learn how Playwright device mapping works, what gets emulated, and the gotchas to watch out for.
A deep dive into custom header injection - context-level vs request-level approaches, Playwright MCP configuration, and when to use each pattern.
Why general-purpose AI agents fail at browser testing, what it takes to make them reliable, and how Test-Lab handles the hard parts.
Relying on one AI provider is a single point of failure. Here's how we built resilience into Test-Lab with OpenRouter and multi-provider fallbacks.
We started with Goose, Block's open-source AI coding agent. Then we built something custom. Here's what drove that decision and what we learned along the way.