Docs

What the bot learns from

Everything the bot can say comes from a source you connected. There is no general world knowledge behind it.

What content can I connect?

  • A website address. Grounly crawls it and re-crawls weekly, so published changes reach the bot without you doing anything.
  • Pasted text. For things that live in nobody's CMS — shipping rules, opening hours, an internal policy.
  • An uploaded file. PDF, plain text or Markdown, up to 10 MB.
  • A Google Drive folder. Docs, Sheets and PDFs, kept in sync by the same weekly sweep.

How much of my website gets indexed?

Your plan sets the page budget per source. On top of that there is a hard ceiling that no plan lifts — see plans and limits. Both exist for the same two reasons. Indexing costs money per page, and a knowledge base stuffed with four thousand near-identical pages retrieves worse than one with two hundred good ones.

What does Grounly see in my Google Drive?

Only the files you click. The connection uses Google's drive.file scope and its own picker dialog, so Grounly cannot list or read anything you did not select — not the rest of the folder, not the rest of your Drive.

The honest cost of that: files added to the folder later have to be selected again. Presentations, images and video are skipped, and the source shows you which files were skipped rather than pretending they were read.

Can I make the bot answer something exactly?

Yes — a custom Q&A pair. It is matched before the model is consulted: first on an exact question match, then semantically above a similarity threshold. On a match your answer is returned word for word and the model is never called. Use it for wording that must be exact, for commitments and legal phrasing, and to override a page that is out of date.

Two things it is not for. It is not a patch for the bot failing to find something — that belongs in a source, or the same gap comes back in a different wording. And it is not translated: a Slovak entry answers an English-speaking visitor in Slovak.

What happens when a page disappears from my site?

A re-crawl does not remove it. Pages that vanish from your sitemap stay indexed until you delete the source itself. If you have retired a product page and it must stop being quoted, delete and re-add the source.

What the chatbot learns from — Grounly