Emotional Intensity
Jonah Berger
High-arousal emotion drives sharing at scale. Castle flags clips that stay flat.
Virality, before you spend.
Know how your video will perform, before you spend a penny.
Everyone uses AI to make the ad now. Almost no one can tell you if it will land. Castle reads the psychology inside your clip and scores it, so you fix the hook before you pay to promote it.
Works with Reels, Shorts, TikTok, LinkedIn — clips up to 60 seconds, and longer videos can be scored on their first 60. Or put two videos head-to-head →
34/100
Won't travelHook Salience (0-1.5s)
Pattern-Interrupt
Emotional Intensity
Curiosity Gap
Concreteness
Narrative Transportation
Social Currency
Practical Value
$0 trillion+
spent on ads every year
0%
of videos now made with AI
0%
lost in ROAS
Analytics tell you that it flopped. Never why.
After-the-fact views and retention charts confirm the miss, but they never name the reason or the fix. Castle reads the clip the way a viewer's brain does, names the exact failing mechanism, and gives you the one change that fixes it.
The fix comes from named research, not a hunch.
Every score names the mechanism behind it. Here is the science each factor draws on, and how Castle puts it to work.
Social Currency
Jonah Berger, STEPPS
People share what makes them look good. Castle checks whether your clip does.
Curiosity Gap
George Loewenstein
An open loop compels action. Castle scores the exact gap your hook leaves open.
Concreteness
Chip and Dan Heath
Concrete detail sticks, abstraction slides off. Castle rewards vivid, specific clips.
Narrative Transportation
Green and Brock
Being pulled into a story lowers resistance. Castle detects the pull and scores it.
Hook Salience (0-1.5s)
Orienting response, Sokolov
The brain decides watch-or-scroll before thought. Castle weights your first 1.5 seconds.
Pattern-Interrupt
Prediction-error, Schultz
Dopamine spikes when prediction breaks. Castle counts every interrupt that re-grabs attention.
Practical Value
Jonah Berger, STEPPS
Useful content gets shared as a favour. Castle checks whether your clip earns it.
Not a score. A diagnosis.
What Castle does
Score, fix, create, launch.
One tool from raw clip to live campaign.
Virality Analyzer
A 0-100 score, a per-second retention curve, and the exact fix for your weakest factor.
Compare with a competitor
Your video vs theirs, head-to-head. Who hooks better, where attention drops, what to copy.
See how audiences react
Six audiences, Gen Z to B2B, plus the emotions your clip triggers. AI-estimated, labeled.
Hook Recommender
Matches your topic to proven hook archetypes, then rewrites your opening line to fit.
One-click Meta launch
Push the winning creative to Meta in one click. Paused-mode, you stay in control.
A confidence score you can check
Every score carries a confidence percentage and its reasons. Nothing invented.
AI content-gen
Generate a clip and on-brand copy so you can ship the fix the same day.
Creative and performance analytics
Track what works across your creatives over time. See which factors lift or kill performance.
Shareable report cards
Send a link. The score, retention curve, and fixes render for anyone, no login required.
The 8 things that decide if your video travels.
Castle scores every one, then shows you the part of the brain it works on. Hover a factor to light it up.
occipital active
Hover a factor to light its region.
Hook Salience (0-1.5s)
Occipital region
0%
The brain commits to watch before you consciously choose.
Factor 1 of 8Pattern-Interrupt
Frontal region
0%
The unexpected re-grabs attention the instant prediction breaks.
Factor 2 of 8Emotional Intensity
Limbic region
0%
High-arousal emotion is what makes people share.
Factor 3 of 8Curiosity Gap
Limbic region
0%
The gap between what you know and want to know creates a felt need.
Factor 4 of 8Concreteness
Hippocampal region
0%
Concrete beats abstract for memory. Vivid detail sticks.
Factor 5 of 8Narrative Transportation
Temporal region
0%
Being carried into a story lowers counter-arguing.
Factor 6 of 8Social Currency
Frontal region
0%
Sharing signals identity. People share what makes them look good.
Factor 7 of 8Practical Value
Temporal region
0%
Useful means shareable. People pass on what helps others.
Factor 8 of 8From link to fix in one pass.
01
Paste or upload
Paste a link or upload your clip. YouTube, or any Reel, Short, TikTok, or LinkedIn video.
02
We read the clip
Castle's multimodal AI samples the frames and audio and scores the 8 factors plus a per-second retention curve.
03
Get the fix
A 0 to 100 score, your weakest factor, and the exact fix, before you spend on promotion.
Compare
Your video vs theirs.
Paste your reel and a competitor's. Castle scores both, shows who hooks better, where each one loses attention, and the exact moves the weaker video should copy. Fresh videos take a few minutes each to score — the comparison is just as thorough either way.
Compare two videosExample head-to-head
Your video
72
Hook 79
+30
Theirs
42
Hook 37
1
Your hook lands at 0.8s — theirs waits until 2.4s.
2
Their clip loses attention at the 9s pacing lull.
3
They should copy your on-screen payoff text.
Dogfooded on Rooted
We use Castle on our own plant-care app's reels. One scored 34 (won't travel). We rewrote the first 1.5 seconds per Castle's fix. It re-scored 89.
See the case study34
before
89
after
Questions, answered.
The clip's structural psychology: hook, pacing, emotion, curiosity. Not your distribution, your audience, or luck.
It is a directional diagnostic grounded in named research. It scores the content's signals, not a guarantee of the outcome.
Paste a YouTube link, or upload any clip: Reels, Shorts, TikTok, LinkedIn.
No. Processing is transient. We keep only the score and the metrics, never the raw clip.
No. Paste or upload and get your score, free.
Yes. Open Compare, add your video and theirs (link, upload, or a past analysis), and Castle shows who hooks better, where attention drops in each, and the moves to copy.
They are AI estimates built from your clip's frames and transcript — how six audiences, from Gen Z to B2B decision-makers, would likely react. Estimated, not measured, and labeled that way in the product.
Stop guessing. Score your next video free.
Score it, see how six audiences react, and put it head-to-head with a competitor's — all before you spend on promotion.