Agentic web census, week 2026-W40
Sealed 2026-10-09. The same frozen sets of the most-visited sites, measured on a schedule from outside: what they tell and serve AI agents this week against last week. No site is named.
top 1,000, daily
Run 3 (finished 2026-10-04), 991 sites, no earlier run to compare with.
| Figure | This week | Change, same sites | Sites +/− |
|---|---|---|---|
| Homepage answers a browser (2xx after redirects) of 991 | 51.5% | — | — |
| Serves a robots.txt (200, not HTML) of 510 | 79% | — | — |
| robots.txt names a sitemap of 403 | 69.7% | — | — |
| robots.txt carries a Content-Signal line of 403 | 3.7% | — | — |
| Content-Signal says ai-train=no of 403 | 2.2% | — | — |
| Content-Signal says ai-input=no or search=no of 403 | 1% | — | — |
| robots.txt closes / to at least one of the 12 AI agents of 403 | 24.6% | — | — |
| robots.txt closes / to GPTBot of 403 | 16.6% | — | — |
| robots.txt closes / to OAI-SearchBot of 403 | 8.2% | — | — |
| robots.txt closes / to ChatGPT-User of 403 | 10.4% | — | — |
| robots.txt closes / to ClaudeBot of 403 | 15.4% | — | — |
| robots.txt closes / to Claude-SearchBot of 403 | 8.4% | — | — |
| robots.txt closes / to PerplexityBot of 403 | 12.4% | — | — |
| robots.txt closes / to Google-Extended of 403 | 13.9% | — | — |
| robots.txt closes / to Applebot-Extended of 403 | 11.9% | — | — |
| robots.txt closes / to CCBot of 403 | 19.6% | — | — |
| robots.txt closes / to Bytespider of 403 | 18.9% | — | — |
| robots.txt closes / to Meta-ExternalAgent of 403 | 15.1% | — | — |
| robots.txt closes / to Amazonbot of 403 | 12.9% | — | — |
| Serves /llms.txt of 510 | 20.6% | — | — |
| Serves /agents.md of 510 | 2.7% | — | — |
| Publishes an agent-skills index of 510 | 2% | — | — |
| Publishes an MCP server card of 510 | 2.4% | — | — |
| Publishes an API catalog (RFC 9727) of 510 | 2.2% | — | — |
| Publishes an A2A agent card of 510 | 0.4% | — | — |
| Declares any agent capability (any of the four above) of 510 | 2.7% | — | — |
| At least one AI identity is refused (non-2xx) where the browser got 2xx of 510 | 17.1% | — | — |
| At least one AI identity gets under 70% of the browser's bytes (both 2xx) of 510 | 3.7% | — | — |
| Names an MCP endpoint (server card, llms.txt, agents.md, a homepage link, or /mcp answers like one) of 0 | —% | — | — |
| An MCP endpoint completes the handshake (initialize and tools/list; no tool called) of 0 | —% | — | — |
| Of sites naming an MCP endpoint, one completes the handshake or asks for credentials of 0 | —% | — | — |
| The A2A agent card names an endpoint url of 0 | —% | — | — |
| Of A2A cards naming a url, the url answers (any non-5xx, not an HTML page) of 0 | —% | — | — |
| Homepage carries WebMCP (navigator.modelContext, a WebMCP bridge, or an mcp+json block) of 0 | —% | — | — |
| GPTBot refused where the browser got 2xx of 510 | 14.7% | — | — |
| ClaudeBot refused where the browser got 2xx of 510 | 13.9% | — | — |
| OAI-SearchBot refused where the browser got 2xx of 510 | 9.4% | — | — |
| PerplexityBot refused where the browser got 2xx of 510 | 10% | — | — |
top 10,000, weekly
Run 1 (finished 2026-10-03), 9776 sites, no earlier run to compare with.
| Figure | This week | Change, same sites | Sites +/− |
|---|---|---|---|
| Homepage answers a browser (2xx after redirects) of 9776 | 57.7% | — | — |
| Serves a robots.txt (200, not HTML) of 5643 | 80.8% | — | — |
| robots.txt names a sitemap of 4562 | 69.7% | — | — |
| robots.txt carries a Content-Signal line of 4562 | 3.2% | — | — |
| Content-Signal says ai-train=no of 4562 | 2% | — | — |
| Content-Signal says ai-input=no or search=no of 4562 | 0.5% | — | — |
| robots.txt closes / to at least one of the 12 AI agents of 4562 | 19.1% | — | — |
| robots.txt closes / to GPTBot of 4562 | 13.7% | — | — |
| robots.txt closes / to OAI-SearchBot of 4562 | 6.4% | — | — |
| robots.txt closes / to ChatGPT-User of 4562 | 8.4% | — | — |
| robots.txt closes / to ClaudeBot of 4562 | 12.3% | — | — |
| robots.txt closes / to Claude-SearchBot of 4562 | 5.9% | — | — |
| robots.txt closes / to PerplexityBot of 4562 | 9.4% | — | — |
| robots.txt closes / to Google-Extended of 4562 | 10.4% | — | — |
| robots.txt closes / to Applebot-Extended of 4562 | 10.1% | — | — |
| robots.txt closes / to CCBot of 4562 | 14.9% | — | — |
| robots.txt closes / to Bytespider of 4562 | 14% | — | — |
| robots.txt closes / to Meta-ExternalAgent of 4562 | 11.1% | — | — |
| robots.txt closes / to Amazonbot of 4562 | 10.2% | — | — |
| Serves /llms.txt of 5643 | 15.4% | — | — |
| Serves /agents.md of 5643 | 1.3% | — | — |
| Publishes an agent-skills index of 5643 | 0.9% | — | — |
| Publishes an MCP server card of 5643 | 0.9% | — | — |
| Publishes an API catalog (RFC 9727) of 5643 | 1.1% | — | — |
| Publishes an A2A agent card of 5643 | 0.3% | — | — |
| Declares any agent capability (any of the four above) of 5643 | 1.3% | — | — |
| At least one AI identity is refused (non-2xx) where the browser got 2xx of 5643 | 22.9% | — | — |
| At least one AI identity gets under 70% of the browser's bytes (both 2xx) of 5643 | 2.4% | — | — |
| Names an MCP endpoint (server card, llms.txt, agents.md, a homepage link, or /mcp answers like one) of 0 | —% | — | — |
| An MCP endpoint completes the handshake (initialize and tools/list; no tool called) of 0 | —% | — | — |
| Of sites naming an MCP endpoint, one completes the handshake or asks for credentials of 0 | —% | — | — |
| The A2A agent card names an endpoint url of 0 | —% | — | — |
| Of A2A cards naming a url, the url answers (any non-5xx, not an HTML page) of 0 | —% | — | — |
| Homepage carries WebMCP (navigator.modelContext, a WebMCP bridge, or an mcp+json block) of 0 | —% | — | — |
| GPTBot refused where the browser got 2xx of 5643 | 19.6% | — | — |
| ClaudeBot refused where the browser got 2xx of 5643 | 19.8% | — | — |
| OAI-SearchBot refused where the browser got 2xx of 5643 | 12.3% | — | — |
| PerplexityBot refused where the browser got 2xx of 5643 | 12.7% | — | — |
Each frame compares the last run that finished in this ISO week (UTC) with the last run that finished before it. Change is like-for-like: points = (sites gained - sites lost) / sites compared, counting only sites measured in both runs and inside the figure's denominator both times, so a site that did not answer in one run cannot move a figure. points_all is the plain difference of the two percentages, and denominator_entered / denominator_left count the sites that moved in or out of a figure's base (often a site that did not answer one run). No site is named; per-site records: https://crawlcheck.io/api/census/records.
Signed JSON: /api/v1/census/weekly?week=2026-W40 · feed: Atom · method: /api/census. Editions: 2026-W41 · 2026-W40.