What source types are
Every page an AI model cites gets classified by the kind of site it lives on: a news outlet, a blog, a Reddit thread, a YouTube video, and so on. The classification matters because the fix for weak visibility depends on where the models look. If your category’s answers are built from Reddit threads, publishing more blog posts won’t move you; if they’re built from news coverage, PR will. Source type is a property of the cited page, so it shows up wherever citations do: as a Source Types filter on Offsite Mentions and on Citations (All Citations and Citation Trends), and as a Source badge on citation table rows and the URL detail sheet. It complements the content type classification (listicle, review, comparison, and so on), which describes what the page is, while source type describes where it is.
The source types Promptwatch tracks
- News: traditional news outlets.
- Blog: blogs and personal sites.
- Corporate: company websites, including yours and your competitors’.
- Academic: universities and research institutions.
- Government: government sites.
- Wiki: Wikipedia and other collaborative references.
- Reddit: posts and comment threads.
- YouTube: videos and their comments.
- X: posts on X (Twitter).
- Facebook, Instagram, TikTok, LinkedIn: posts on each network.
- GitHub: issues, discussions, and READMEs.
- Stack Overflow: developer Q&A.
- Medium and Substack: publishing platforms and newsletters.
- HackerNews: posts and comment threads.
- Quora: Q&A threads.
- Other: everything that doesn’t fit a category above.
The Socials section: Reddit, YouTube, and LinkedIn
Reddit, YouTube, and LinkedIn have dedicated pages under Socials in the sidebar, with a model filter and date range at the top.
- Reddit: a Top Posts Over Time chart with the most-cited posts beside it, and a table of every cited post with its subreddit, upvotes, comments, citation count, average citation position, and post state (archived, locked, or deleted). The subreddit column is the strategic one: it tells you which communities feed the models in your category. Full walkthrough: Reddit citations.
- YouTube: the most-cited videos and channels, overview counts (total YouTube citations, unique videos, distinct channels), and a table of every cited video with channel, views, likes, comments, citation count, average position, and duration. Channels that keep getting cited are collaboration and sponsorship candidates. Full walkthrough: YouTube citations.
- LinkedIn: every cited LinkedIn post with its author, reactions, comments, citation count, average citation position, and date. In B2B categories this is where the influential voices live, and the author column names them. Filter by prompt, topic, or tag, and export the table like any other citation view. Full walkthrough: LinkedIn citations.
My channels
To focus on the accounts your brand runs, open Brand → Channels, choose a platform, enter a URL or handle, and click Add. You can save a Reddit subreddit or user, a YouTube channel, or a LinkedIn company page, showcase page, or profile.
