Mentions
One feed for every platform. This is where you will spend almost all of your time in BrandSonar.
Mentions is the first item in the sidebar and the app's home screen. It shows everything BrandSonar has found across Reddit, X, YouTube, Hacker News, and Bluesky in a single stream, newest first. Scroll to load more; the feed also refreshes itself about once a minute, so you can leave it open.
What is on a mention
Each card packs the whole judgement into one row:
- Platform and author
- The source icon and the account that posted, plus the subreddit or channel where that applies.
- Highlighted text
- Your keywords are marked in the body so you can see at a glance why the mention matched. Long posts clamp to seven lines with a Show more control.
- Relevance dot
- The coloured dot on the right is the relevance tier: green, amber, or grey. Hover it for the exact score.
- Sentiment
- The face icon beside it: positive, neutral, or negative.
- Labels
- Applied automatically by AI. Use the tag icon in the footer to add or remove them by hand.
- Original post
- The arrow at the far right opens the mention on the platform it came from.
Filtering the feed
Two filters are always on screen, Relevance and Period, and + Filter adds the rest: platform, sentiment, label, and keyword. Each filter becomes a chip you can change or dismiss.
There is no free-text search. Filtering by keyword narrows the feed to mentions containing one of the terms you already track, which covers most of what a search box would be used for.
Relevance scores
Every mention is scored from 0 to 100 for how much it is actually about your brand, and the feed groups those scores into three tiers. The Relevance filter sets the floor:
- High
- 80 and above. Directly about you.
- Medium
- 65 and above. The default. Likely relevant, sometimes a passing reference.
- Low
- Everything scored. Includes loose matches and posts that merely share a word with your brand name.
If too much noise is getting through, reach for negative keywords before raising the relevance floor. Negative keywords remove a specific kind of false positive; a higher floor removes everything marginal, including the edge cases worth seeing.
Sentiment
Each mention is classified positive, neutral, or negative. It earns its keep in two places: as a filter, where negative mentions in the last seven days makes a short, high-value list, and as a trend on the Analytics page.
Labels
Labels are your own categories, applied to incoming mentions by AI. Clicking the tag icon on a mention opens the full list so you can correct a call or add one the model missed.
Competitor mentions
Competitors have their own feeds. Open one from the Competitors list in the sidebar to see just that brand, or use the combined competitor view to read all of them together, with a chip on each card showing which brand it belongs to.
Exporting
Export in the top right downloads the feed as CSV, XLSX, JSON, or Markdown. The export has its own filters for platform, date range, minimum relevance, sentiment, and label, pre-filled from whatever the feed is currently showing.
Exports are available on paid plans. For a programmatic feed rather than a file, use the REST API.