nectr

Your pages as markdown and data, the way AI agents read them

nectr turns a web page into clean markdown, including what loads with JavaScript, and shows what an AI agent gets from it. Scan is open to everyone. Integrate, Publish and Monitor are coming: a markdown version of every page on your site, its data as datasets, and who reads them.

nectr turns a web page into clean markdown: the whole page, including what loads with JavaScript. It is the version of a page that AI agents and other tools can read in full, without menus, scripts or markup.

Scan: Enter a page's URL. With an account you get:

  • the whole page as markdown, scripted parts included;
  • what only appears with JavaScript, and its share of the page's text;
  • what the site's robots.txt allows the main AI crawlers, such as GPTBot, ClaudeBot and PerplexityBot;
  • whether up to three PDFs the page links to have text, or are only scanned images;
  • the estimated tokens of the markdown, and of the raw HTML for an agent that reads it as it is;
  • the page without JavaScript: the text in its HTML.
  • scans are also available through the API, without a key but without full markup. Scan a page.

How nectr works

  • nectr's crawler is Nectrbot, and it follows robots.txt.
  • A scan's result goes back to you only. nectr keeps a daily count of scans, with no address or result in it.
  • nectr doesn't train AI models or sell content.
  • nectr runs in Switzerland, with Infomaniak.

Coming next: For sites whose owner has verified the domain:

  • Integrate: a markdown version of every page and PDF on your site, refreshed every night and linked from the page itself, with an llms.txt and the last 10 changes to each page.
  • Publish: a page's tables and lists, shared as datasets (CSV or JSON).
  • Monitor: which AI agents read your pages, a weekly email, and an alert when a page loses text.

Questions? Get in touch.