Perspective
The verification layer for your software factory
Agents industrialised production but not proof. Four layers of a delivery pipeline, why review and tests structurally cannot cover the third, and what the missing one has to be.
Blog
Notes on browser QA for AI-speed teams.
Practical writing on agentic testing, MCP workflows, browser automation, and keeping product quality visible as teams ship faster.
Guides
Longer, evergreen pieces on how Kery fits next to the tools you already run.
Setup
Claude Code + Playwright MCP for your PRs
A working setup: what to install, the four inputs it needs, the prompt that makes it verify rather than summarise, and the auth step that stops most attempts.
Comparison
Claude Code + Playwright MCP vs. Kery
The DIY agent-drives-a-browser setup: where it genuinely holds up, and the six things it cannot do.
Comparison
Playwright, without a suite to maintain
Nobody abandons a test suite because writing it was hard. They abandon it in month four. Where derived checks fit instead.
Comparison
Automating the repetitive half of QA
What to hand to a machine, what to keep human, and why smoke testing stopped keeping pace with agent-authored pull requests.
Your stack
Verifying what Claude Code ships
Write with the agent, verify with something that did not write the code. Plus running Kery from inside the session over MCP.
Your stack
End-to-end testing for Cursor
Background agents open pull requests nobody watched happen. Two ways to wire a browser check into that loop.
Your stack
Testing every Next.js preview deploy
Vercel already builds a preview for each pull request. Where App Router apps actually break once something opens it.
Use case
A browser pass on every pull request
What lands on the PR, why the bar for failing a build is deliberately high, and what the evidence looks like a week later.
Use case
Testing the flows behind your login
Clerk, Supabase, Auth0, Firebase, 2FA codes and magic links — and why storage-state fixtures keep going stale.