Hey Product Hunt! I'm Ashraf, the builder behind Mutatr. When I was an applied scientist at Microsoft, experimentation was a core part of the job; but the process was beyond painful. Countless stakeholders, complex tooling, and weeks of setup before a single test could run. If it's this painful in big tech, small teams with little-to-no traffic have no chance. So I wondered: what if an AI agent just handled the whole thing? You give Mutatr a page and a goal. It suggests experiments, writes the variant code in your actual codebase, renders it, and scores it against synthetic customer personas — complete with attention heatmaps and a scorecard telling you whether to ship, iterate, or stop. The whole loop runs in minutes, not weeks. No SDK to install. No traffic needed. Under the hood it's Claude Agent SDK + Playwright, running locally on your machine. Your code never leaves your desktop. We're launching with Next.js support first — more frameworks coming soon. Would love your feedback. What would make this useful for your workflow?

Mutatr

Agent that builds & scores your A/B variants in minutes

Agent that builds & scores your A/B variants in minutes