Human · desktop · Chrome00:00
❯ lands on / from a search
└ Largest paint lands at 3.6s. Load scores 75 against a 90 benchmark.
└ load 75 · benchmark 90
✓ ranked · Improve load speed to the 90 benchmark
TESTPLAN · FREE · NO ACCOUNT
Paste a page. In a few minutes, PlayHero reads it the way a buyer does and the way an AI agent does, then ranks your top three A/B experiments by what's most likely to move pipeline.
3 EXPERIMENTS FREEA FEW MINUTESNO ACCOUNT
Human UX · Agent UX
A buyer skims your hero on a phone. An AI agent fetches your page to answer a question about you. PlayHero scores both readings across seven pillars, and Testplan ranks the fixes that move both.
How fast the page paints, whether the hero says one thing, whether there's a way to start without talking to sales. The findings that became this page's top three.
What an answer engine, a shopping assistant or a crawler can read without a browser. The checks the agent lens runs on every scan.
FULL PLAN · 2 CREDITS
Unlock the complete ranked roadmap with ICE scores, supporting evidence, run-ready hypotheses, and exports. Credits are used when you generate the plan, not when you export it.
HOW IT WORKS
Simple enough to start now, structured enough to hand off.
Point Testplan at any URL. PlayHero reads the page's structure, user experience, and conversion signals.
> playhero scan your-site.com --testplan
> 14 findings · 7 pillars · ranking…Findings become experiment candidates, ordered by what's most likely to move pipeline, not by what's loudest in the backlog.
01 · Improve load speed to the 90 benchmark
Pillar: LOADTake the top idea into standup today, or generate the full plan with evidence, hypotheses, and exports ready for Jira or Linear.
HYPOTHESIS: We believe a sticky CTA will lift trial starts… Metric: trial-start rate · MDE 12%
FAQ
What Testplan does, where it stops, and what your team controls.
The top 3 ranked experiments on the spot, no account. The results page unlocks up to 5.
2 credits generate the full ranked plan — ICE scores, evidence, and hypotheses. Copying it as Markdown, Jira wiki, or Linear markdown costs nothing extra.
No. It ranks and hands off. You run the test in your own tool, and can declare pre-registration fields in the app before you start.
Rallycards from your scan, competitor gaps from Compare, and multi-page SuperScan findings when you've run one.
For signed-in full plans, yes. Each one is checked against your earlier plans for the same site and repeats are marked. Free runs aren't checked; there's no account to remember them for.
YOUR NEXT TEST, RANKED
Get your top three free in a few minutes, or generate the full ranked plan when the team needs evidence, ICE scores, and exports.
3 EXPERIMENTS FREEA FEW MINUTESNO ACCOUNT
Nothing is declared yet. Declaring the primary metric, minimum detectable effect, stop rule and maximum duration is what locks the item into a test.