Mention inventory
Filter stored mentions by sentiment, domain, and date. Open validated source links or export the rows already loaded.
Brand Radar
Scan the conversation Keep the evidence.
Run a scan when you need one. Review mentions, sentiment, top domains, scan-over-scan change, and a digest that cites the stored mention rows behind it.
From mention to evidence
One scan, with the source still attached.
The workspace keeps normalized mention rows, deterministic aggregates, and cited interpretation distinct.
Deterministic totals
Read mention counts, a four-way sentiment distribution, and top citing domains computed from retained rows.
Scan history
Compare settled scans for the same query. A first scan stays “no comparison,” never a fabricated zero.
Cited digest
Digest sentences are labelled AI interpretation and expose their stored-row citations. Unreliable candidates are dropped.
Product truth
Illustrative output, labelled by provenance.
This frame mirrors the shape of a stored result. It is not a live account, screenshot, or monitoring claim.
| Signal | What the product shows | Source |
|---|---|---|
| Mention row | Bounded title, snippet, domain, and observation time | Provider observation |
| Sentiment split | Positive, neutral, negative, and unknown | Deterministic aggregate |
| Digest sentence | Displayed only with stored-row citations | AI interpretation |
Honest limits
A scan is a scan—not a monitor.
Cadence, source, and partial-result limits stay visible before you act on the output.
On demand only
Brand Radar does not continuously monitor the web or schedule scans. Weekly Pulse can read stored scans; it never starts one.
Evidence from this run
A scan reflects normalized provider observations retained for that run, not a complete census of every mention online.
Partial stays partial
A budget ceiling or provider interruption can leave useful retained rows. We keep that evidence and do not invent a more specific stopped stage.
On-demand intelligence
Run the scan when the question matters.
Start with an account, or audit the site first and see what deserves attention.