Turn a website visit into inspectable evidence: requests, cookies, browser signals, and consent behavior.
Static research dashboard tracking public signals for the dead internet theory: AI content, AI bot traffic, Common Crawl samples and human web activity.
Reproducible HTTP-layer measurement of how the most-visited Turkish and European websites answer AI crawlers — with an open dataset.
DYNAMICS-WEB: a measured model of the character of web content, propagated across the link graph. Open paper (Zenodo 10.5281/zenodo.20981883), CC BY 4.0. Content sibling to DYNAMICS-8.
How much of the web can AI assistants actually read? Daily measurement of every crawler, domain by domain.
Scans websites for hidden text aimed at AI agents. Powers a reproducible prevalence study.
Kontrollierte Messung des google.com/goto-Link-Wrappings auf deutschen Google-Suchergebnisseiten. Rohdaten, Messprotokolle und Auswertungsskripte, erhoben am 2. September 2026.
Uptake — publishing for machines that copy. A field manual for the web after search, built on 14 days of repository traffic: 961 clones from 501 distinct machines against 36 page views. Text and data CC BY-SA 4.0.
Which AI discovery routes exist on the web, how widely they are adopted, and whether they can actually be fetched. Measured, not estimated.
A census of the agent-reachable web: can an AI agent discover and connect to an MCP server for a given domain? Open dataset, open code.
To associate your repository with the web-measurement topic, visit your repo's landing page and select "manage topics."