Skip to main content
Plans: All plans · In the app: Sidebar → Sources → Source Gaps
A source gap is a page that AI engines cite in their answers, that names one or more of your competitors, and that never names you. Engines learn from those pages, so every answer that leans on one hears about your competitors instead of you. Getting your brand onto them is one of the most direct ways to change what engines say. In the sidebar, open Sources. Source Gaps is the first page in the section, above Page Tracker, Citations and Social. It belongs to the brand you have selected.

How it works

1

Answers cite pages

Every tracked answer lists the pages it used, with the engine and the date.
2

QuotableBot reads the cited pages

Pages cited in the last 30 days are queued and read in the background, most cited first. QuotableBot keeps the readable text of each page. No AI model reads the page.
3

Brand names are matched in the text

Quotable looks for your brand and each tracked competitor in the page’s own text: whole words, ignoring accents, including the short forms a multi-word name goes by. Brand names that are also everyday words are guarded, so “notion” in a sentence does not count as a mention of Notion.
4

Gaps are picked out

A read page is a gap when it names at least one competitor and does not name you. Your own site never counts. A competitor named on its own site, or a subdomain of it, does not count either: no outreach gets you onto a competitor’s product page.
Source Gaps uses the same cited pages and the same rule as the Names rivals, not you filter on Citations, so a gap here is always a row there.

Gap score

Each gap gets a Gap score: the number of competitors the page names, multiplied by the times answers cited it in the date range. A page naming three competitors and cited 20 times scores 60. A domain’s score is the sum of its gap pages’ scores. Start at the top: a high score means many answers, and many competitors, ride on that page.

Headline tiles

The tiles size the whole problem. They count every gap, not only the rows your filters leave.

The Gaps table

The table opens grouped by domain. Switch to one row per page with Group by in the Filters menu. Your choice is kept in the address, so a reload or a shared link opens the same view. The table shows 12 rows at a time, sorted by Gap score. Click any column header with a sort mark to sort by it.
Click a domain to expand its gap pages under it, highest score first.
Click a page to open its cited page view. The arrow next to the title opens the page itself in a new tab.

Filters

The page’s Filters menu holds the usual view filters (models, date range, topics, locations, competitors and prompt type) and these for the table: Search pages matches the page address, its title and its tags. Clear filters appears when nothing matches.

Bookmarks and tags

Bookmark a page or a whole domain to build a shortlist: hover the row and click the bookmark. Add tags such as outreach or q3 from the Tags button on the cited page view. A page and its domain each take up to 12 tags of up to 32 characters. Tags and bookmarks are shared with everyone on the brand. Viewers can see them but not change them.

Export to CSV

CSV downloads the rows your filters and search leave, in the current sort, across all table pages.

Open a gap

Clicking a page opens its cited page view, with a link back to Source Gaps. It shows citations over time and by model, the prompts and answers that cite it, and every tracked brand the text names. A banner says which competitors the page names while leaving you out. Create content with the agent in that banner opens Quotable Agent with a request already written: who to contact, a short outreach pitch, and what you could publish to earn the mention. For this list every week, start a routine from the Source gap sweep template.

Freshness and limits

  • Reading takes time. Pages are read in the background, most cited first. A new brand shows Reading the cited pages until the first ones arrive.
  • Pages are re-read about every 7 days. When the text changes, the previous version is kept. On the cited page view, Read again reads a page on demand, at most once an hour.
  • Some pages can’t be read, for example pages behind bot protection or on sites whose robots.txt refuses QuotableBot. They are marked as unreadable, with the reason. They are never counted as “doesn’t name you”, so they never become gaps. QuotableBot respects robots.txt; see QuotableBot.
  • Competitor changes apply without re-reading. Add or rename a competitor and pages already read are matched again from their stored text.
  • You need competitors. With none tracked, the page asks you to add them in Settings. See Brands and competitors.
  • Only the most cited pages count. Gaps are found among the 2,000 most cited pages in the date range.
  • A mention is a written name. A page that links to you or shows your logo without writing your name still counts as a gap.

Next steps

Citations

Every page and domain AI engines cite.

Page Tracker

Which of your own pages feed AI answers.

Routines

Get the gap list every week.
Last modified on November 18, 2026