The Sources page Query Type filter (Branded / Open-Ended / All) used to only affect the table. It now drives the entire page — the summary cards, top-domains dropdown, and domain rollup all recompute when you switch modes. Filtering to Branded is particularly useful for tracing the path AI agents take when answering brand-specific questions. The filter is also now a URL parameter (?queryType=), so you can bookmark or share a filtered view and it opens correctly. The page defaults to All Queries; previously it defaulted to Open-Ended, which left the summary cards showing different data than the table.
The Visibility Trends graph has a new Priority grouping option alongside Topic, Prompt, Persona, and Platform. If your prompts are tagged with reach levels — Core, Adjacent, Aspirational, Visionary — you can now chart mention-rate trends by those levels over time. The toggle only appears when that data exists, and the priority filter hides automatically while you're using the grouping to avoid redundancy.
The "How scores work" modal on the Topics page now links out to the full methodology doc, and the modal itself is accessible directly from the score card's "Latest Report" link — previously you had to find it elsewhere in the UI.
The Yelp ads channel is now tagged Roadmap rather than showing a live badge. Yelp campaigns are simulated, so the green indicator was inaccurate. The card and demo flow are still there — the label is just honest about the status.
The prompts table on the Topics page now displays correctly. Columns were collapsing and wrapping text to a few characters wide; the table now uses fixed proportional widths, gives the Prompt column the most space, and scrolls horizontally on narrow screens. The research hub index page (/research) now appears in the sitemap and has a footer link from every page — previously 123 of 130 published articles were indexed but the hub itself was missing from sitemap.xml and had no consistent inbound link.