Detection did not find my table or list
Detection ranks the tables and lists it finds and marks the best match. If the one you want is not there, or it has the right items but the wrong columns, choose Pick on page: click two example items, then the values inside one of them that should become columns. The pick guide walks through it.
Detection also looks only when you ask. Nothing is read when a page loads, so a list that appears after you scroll or click on the page is found by pressing Scrape again once it is there.
Nothing happens when I press Scrape
The first time you scrape, the browser asks to let Rowspell read the sites you visit. It asks for all sites at once, because a run reads whichever page you are on; allowing it for specific sites only is not enough. If you said no, the panel’s own button asks again. Nothing is read until you allow it.
Chrome 137 or newer is needed, or an Edge built on the same engine.
The count says “4,000+” — is that all of it?
No: a plus sign means at least that many. On a very large page, detection stops counting before the end, so the number is a floor. The panel also says how many rows a run keeps at most: 10,000 on the free tier.
All pages stopped before the last page
All pages clicks the site’s own Next link, numbered pager or “Next” button until there is none. When it stops for any other reason, the done screen says why:
| What the done screen says | What happened | What to do |
|---|---|---|
| Nothing — it just finished | There was no Next control left: the natural end | Nothing; that was every page |
| The page limit was reached | A free run reads up to 50 pages | Continue from page N, in the dock or in History |
| Two pages in a row added no rows | The site kept showing the same items | Check the last page in the browser; the rows you have are kept |
| A page failed or was lost | The site refused or did not load a page | Run it again later; if History offers Continue from page N, use that |
A page that gave no new rows, or that the site refused, does not count against the limit. Every row collected exports, whatever stopped the run.
Scroll and collect ended early
Scroll and collect keeps scrolling while new rows arrive. It stops when the feed stops loading, or when its time budget runs out, and it tells you which: “ended early” means the feed stopped before reaching its bottom. Scroll a little in the browser to wake the feed, then run it again.
A banking, health or government page asks me to type a confirmation
That is on purpose. These pages are read only after you type the confirmation. URL lists and detail crawls refuse them, including a page that redirects to one, and the AI help refuses them before anything is sent.
A page I am signed in to shows a login screen
Rowspell reads the page as your browser shows it, so a site you are signed in to works, and your sign-in details are never read. If the page in the tab shows a login form, sign in there first, then press Scrape.
I closed the panel and the rows are gone
Rows you scrape are held in memory only, by design: closing the panel clears them. Export or copy before you close it. Saved recipes, run history and settings stay.
The site blocks it or shows a CAPTCHA
Rowspell does not solve CAPTCHAs or get around a site’s protections. If a page cannot be read, the panel says so rather than handing you an empty file.
Try it on a page made for it
The practice page has a plain table and a product list across three pages behind a Next link. If a scrape works there and not on your site, the difference is the site.