Today we're flipping the switch on Cortex 1.0.
Cortex predicts how a paid-social ad will perform against your brand's own reference library — using a model that maps video and audio into predicted neural activity. The model is the same one we've been refining for the past year; what shipped this week is the full product around it: a complete brand-bank workflow, real-time candidate scoring, billing with overage, team workspaces, an audit log, and the design rebuild that brings the whole thing into one clean dark-mode interface.
What's in the launch
- A scoring you can read in 30 seconds. Every result page leads with a calibrated winner-probability, then a three-score breakdown, then the brain-network timeline and divergence heatmap that explain the call.
- Per-brand signatures, not a generic model. Upload your historical winners and losers; Cortex builds a brain-response fingerprint specific to your brand. Every score is against your bank, not against a universal benchmark.
- Bank diagnostics + calibration. Find mislabeled references before they pollute scoring. See your signature's leave-one-out winner/loser distributions and the optimal threshold for your brand.
- Real billing, real teams, real audit. Stripe metering against monthly caps with overage, role-based workspaces, SAML SSO on Growth and above, full audit log on Agency.
- A pixel-perfect rebuild. The interface you're reading this on is the new one — every screen redesigned around the brain-signal language that defines what Cortex does.
What's next
We're spending the next six weeks listening. The product is in the hands of customers, and the next set of features (per-second comparison between two ads, on-prem deployments, a public API) lands based on what you ask for.
Try it: start a 14-day trial — no card, 5 free scorings, cancel anytime.