About

Who runs this site, and how it works

WebMCP Agency is a research site about the agent-ready web. It is published by Convertlyft, the web analytics product, and written by the Convertlyft team.

What we do, and what we do not

• We publish guides to WebMCP, llms.txt and AI crawlers, written from primary sources: the WebMCP draft, Chrome's documentation, and each AI vendor's own pages. Every guide lists its sources and the date we read them.
• We run a free checker and a monthly crawl of 10,000 popular US web origins.
• We do not sell services. Despite the name, this is not an agency you can hire. If you want the full report on your site, or to keep watching it, that is Convertlyft.

The crawler

Both the checker and the research crawl run on the crawler behind Convertlyft. It identifies itself as ConvertlyftBot in your server logs. It obeys robots.txt, never calls a WebMCP tool and never submits a form. The one exception to identifying itself: to measure whether your server really admits each AI agent your robots.txt allows, it requests a page using that agent's name.

How the research is done

• Sample: the Chrome UX Report's list of popular origins for US Chrome users, top 10,000 in popularity bands (1k, 5k, 10k), published monthly as CrUX top lists.
• Depth: each site's homepage and one linked page, loaded in Chrome so tools registered by scripts count too.
• What we keep: one summary row per site (states, counts and tool names). The pages themselves are deleted after each site is summarised.
• What we publish: totals, each with the number of sites it was measured over. SEO findings use only issue types the crawler's own test harness verifies.
• Names: we never name a site for a problem. The directory names only sites whose WebMCP tools work, and any listed site can ask to be removed.
• No customer data: the research never uses data from Convertlyft customers.

Contact

Corrections, removal requests and questions: support@convertlyft.com.