We build discourse-intelligence systems that measure public narratives in real time — so open societies see the move forming.
Spot narratives as they take shape across public feeds — before they consolidate into consensus.
Cross-account behaviour computed at the cluster level. Inauthenticity is judged, never assigned per account.
Group-level segment mapping of who a narrative moves — a complement to survey research, not a replacement.
Measure how far a narrative travels and whose discourse it shifts — always at the group level.
Public discourse is captured and structured. Nothing is inferred that was not first observed.
Volume, velocity and framing are computed. Meaning is read and narrated by the model.
Cluster-level behaviour reveals coordination. Inauthenticity is judged, never per account.
The output is early warning — a complement to survey research, not a forecast.
Public discourse is watched around the clock, in the languages of the region.
Every psychographic and coordination read stays at the group or cluster level.
No per-account surveillance. No individual profiling. Ever.
A six-week study of how a cross-border narrative formed and spread across public Facebook communities, mapping the clusters that carried it and the audiences it moved.
We connect the public sources where discourse lives into one measured system — no manual collection, no blind spots.
Code counts, the model narrates. We never claim what we did not observe.
Psychographics and coordination stay at the group or cluster level — never per person.
Sentra measures discourse as it moves — alongside survey research, not instead of it.
Early warning from observed signals. We surface the move — we do not forecast it.
Sentra saw the coordination forming days before it surfaced in our coverage — and showed its working the whole way.
Based on partner
evaluations