Collect brand mentions for any period right in the chat

The agent works with the Brand Analytics Statistical API using the token from your account: first it reads the list of monitoring topics, then for the chosen topic it collects the mentions themselves with full text, daily statistics, sentiment distribution, a breakdown by source type, geography and the top authors by reach. Out of that come the daily negativity dynamics by platform and the picture of where the brand is talked about most.

The key trait of the service is that everything is tied to topics. Mentions do not exist on their own: their boundaries are set by the topic search query configured in the account, and the agent sees exactly what that query captured. So it separately compares topics against each other: it finds duplicates by object and by query wording and shows the topics that sit idle while still burning the plan quota.

The agent never writes here — the Statistical API is used read-only. It neither creates nor edits topics, search queries or filters: if a selection picks up noise or misses what you need, the agent names the specific topic and what should change in its query, and you make the edit in the account yourself. Everything else — collecting, analysing and consolidating mentions — stays on the agent.

Access comes from a single API token issued in the account: it is attached to every request, while the volume you can pull is capped by the plan quota, which is shared across all topics. The agent spends it carefully — it fetches the period in portions, does not re-request what it has already parsed and holds the request rate itself. If a topic has no mentions for the period, it says so directly instead of passing an empty table off as an absence of negativity.