We built Refute because manipulation travels faster than trust teams
Our founders came from media intelligence and machine learning backgrounds, watching brand safety teams and newsroom trust desks repeatedly blindsided by coordinated fake content that looked organic until the damage was done. Founded in London in 2024.
Give every brand safety team and newsroom trust desk the detection capability that only the largest platforms have built internally
Platform trust teams at large social networks have invested years and significant headcount into coordinated inauthentic behaviour detection. That capability sits behind proprietary systems and is unavailable to the practitioners most affected by its absence: mid-market brand safety analysts and independent newsroom editors making real-time decisions without an OSINT team to call.
Refute makes that detection layer accessible via API and a straightforward scoring interface. No research team required. No six-figure contract. Just a scored verdict with the evidence to act on it.
The team
Built content-integrity tooling for media organisations across several years. Deep background in trust-and-safety systems and editorial verification workflows. Saw first-hand how brand safety teams lost hours to manual account review during manipulation campaigns.
Machine learning and natural-language processing specialist with years of experience building trust-and-safety classification systems at a large social platform. Leads product direction and the core scoring models that power Refute's detection pipeline.
Years studying coordinated inauthentic behaviour campaigns across social platforms. Academic background in computational social science and network analysis. Authors the methodological research that informs Refute's detection thresholds and scoring criteria.
Contact us
We respond to every message within one business day. For enterprise enquiries and custom deployment discussions, use the demo booking form.
Book a Demo