CURNT ingests creator content in real time, scores every post across four signal layers, and turns raw social output into structured decision-support for brands.
Three stages. Every post from every verified creator flows through them within seconds of going live.
We continuously sync content across Instagram, TikTok and YouTube. Public data for discovery, verified data via authenticated pulls when a creator connects.
Each post runs through the CURNT engine. Performance scoring, theme extraction (Claude), sentiment analysis, brand-fit prediction against active briefs.
Brands query the engine with a brief. Get ranked creators, ranked posts, and structured reasoning behind each match — explainable, not black-box.
The CURNT Score is not follower count. It's a composite of four independent signal layers, each trained on what actually moves campaign outcomes.
Raw performance normalized by category, platform and creator baseline. Not vanity metrics — the signals that predict campaign lift.
AI-driven theme extraction across every post. We map what a creator actually makes — routines, reviews, city guides, workouts — not what a hashtag says.
Deep comment analysis. Emoji patterns. Response depth. Whether the audience is buying — or just scrolling past with a heart tap.
Content-level alignment between what a creator makes and what a brand wants. Trained on live campaigns — sharpens every week.
Every score is built from a stack of data with different provenance. Brands see which signal came from where.
Scraped and enriched from Instagram, TikTok and YouTube. Follower counts, public engagement, category signals, growth. Enough to discover — never enough to decide.
Real analytics from creators who connect their accounts. Audience demographics, story reach, save rates, saves-to-reach ratios. Ground truth public data can't give you.
Post-campaign performance across our network. Which creator delivered on brief. At what price. For which brand. This is how the model gets sharper every week.
Brands describe the campaign in natural language. The engine responds with matches, scores and reasoning — usable in minutes, not weeks.
Natural language brief. Audience, category, budget, tone. Two minutes. No structured forms — the engine parses what matters.
Each creator scored across all four signal layers. Full reasoning per match. Filtered to your budget range. Ready to reach out — you keep the relationship.
We're onboarding brands into the pilot. Get access to the query interface and start running briefs through CURNT.
Request Demo