Hey Product Hunt — I built isWebMCP because the web is becoming agent-accessible faster than it is becoming agent-verifiable. A page can mention WebMCP, expose a tool, or look easy to automate and still fail the task that matters. I wanted a way to separate four very different questions: - What can we infer from bounded public source? - Is the tool contract narrow, understandable, and safe? - Did an agent actually complete and verify the task? - Was that path measurably better than using the visible UI? isWebMCP turns those questions into an evidence ladder. You can scan a public URL, lint a sanitized tool inventory, explore 24 interactive before-and-after patterns, and run the same synthetic task through UI and WebMCP paths. Evidence keeps its provenance. Missing runtime data stays unknown. Lift is withheld unless the comparison qualifies. We also published the uncomfortable part of our own crawl: WRI v1 scheduled 100,000 popularity ranks, but an integrity audit found two scanner-wide failure windows. The public artifact is therefore labeled audited-partial and uncalibrated, with 34,620 infrastructure errors quarantined rather than blamed on websites. WebMCP itself is experimental Community Group work, and this is an independent project—not an official standards or browser product. I would especially value feedback on one question: **what evidence would you need before trusting an “agent-ready” claim for your own web app?** Try it: https://iswebmcp.com

isWebMCP

Turn UI guesswork into actions you can prove using WebMCP

Turn UI guesswork into actions you can prove using WebMCP