Why we're not adopting Passmark

tl;dr — Passmark is a Playwright wrapper with an FSL license and 185 stars. Stagehand does the same thing under MIT with a larger community. We’re evaluating Stagehand instead.

What Passmark offers

Passmark promises AI-driven browser testing by wrapping Playwright with natural language selectors. Sounds great for verification — let AI describe what to check instead of writing CSS selectors.

Why we passed

It wraps Playwright. Passmark isn’t a Playwright replacement — it’s a layer on top. If we’re already going to depend on Playwright, the wrapper needs to justify its existence.

FSL license. The “Competing Use” clause is a non-starter for a verification tool. If someone uses Maina to verify Passmark’s competitors, we’d be in legal gray area.

185 stars. Not battle-tested. For a tool that gates whether code merges, we need reliability data.

What we’re doing instead

Stagehand (MIT, larger community) does the same thing — AI-driven browser automation on top of Playwright. We’re running a feature-flagged evaluation with hard go/no-go criteria: pass rate, token cost, cache hit rate.

Decision record: ADR 0018 — No Passmark