Blog
Engineering notes
How we build — and measure — an auditable multi-model fact-checker.
Where checkable claims hide (and how we taught our extractor to find them)
AI extractors reliably catch a sentence's headline fact — and quietly miss the claims tucked into appositives, asides, and subordinate clauses. How we measured that gap, closed it, and made every claim traceable to its exact source text. With an interactive demo.
What multi-model consensus catches that single models miss
Any one AI model fails quietly and confidently. A panel of rivals, an adversarial challenger, and an independent judge fail loudly — here's the mechanism, failure mode by failure mode.
How we benchmark a fact-checker (and why our numbers aren't up yet)
Accuracy claims from AI products are usually unfalsifiable. Here's the evaluation we've designed instead — the datasets, the metrics, and why the full suite runs against our October 2026 quality milestone.