Analyzer · 100% free, no signup · Updated May 2026
Score transcript chunks on 5 viral patterns.
A free analyzer that scores a transcript segment on the five evidence-based viral patterns: curiosity gap, contrarian take, emotional peak, data drop, and story hook. Produces a 0-100 composite score plus a verdict (ship-it / decent / weak) and a targeted rewrite hint for the weakest dimension.
Paste a 50-500 word transcript chunk. The tester scores it 0-100 across 5 viral patterns: curiosity gap, contrarian take, emotional peak, data drop, story hook. The composite score is the average of the 5.
Paste transcript chunk
Composite score
45/100
Decent
Curiosity
25
Contrarian
33
Emotion
0
Data
100
Story
67
Strengthen weakest pattern: emotional
Add stakes or feeling: name what was at risk, what was lost, what changed.
Who this is for
Podcasters and video creators scanning long transcripts for the moments that'll actually go viral as shorts.
The problem this solves
Most clip-selection happens by gut feel. The tester encodes 5 evidence-based viral patterns (curiosity gap, contrarian take, emotional peak, data drop, story hook) into a fast scoring engine so creators stop second-guessing.
The five most predictive viral patterns in short-form video are curiosity gap, contrarian take, emotional peak, data drop, and story hook — segments scoring 65+ on the composite outperform random clips by 4-7× on retention.
Source: Stanford-led analysis of 1.2M TikTok clips (2024) + internal Clipflow A/B testing across 8,400 podcaster-published clips
How to use it
- 01
Paste chunk
50-500 word segment from your transcript.
- 02
See per-pattern scores
0-100 on curiosity gap, contrarian, emotional peak, data drop, story hook.
- 03
Composite + verdict
Average score + verdict: ship-it / decent / weak.
- 04
Pattern tips
For your lowest-scoring pattern, get 1-2 rewrite suggestions.
What you get
- ✓ Composite score 0-100
- ✓ 5 viral-pattern sub-scores
- ✓ Verdict: ship-it / decent / weak
- ✓ Rewrite hint for weakest pattern
Frequently asked
What are the 5 viral patterns this tool tests for?
Curiosity gap (unresolved question), contrarian take (against conventional wisdom), emotional peak (high-affect language), data drop (specific number/statistic), and story hook (past-tense pivotal moment). Each predicts retention independently.
How long should the transcript chunk be?
50-500 words — roughly 15-90 seconds of speech. Shorter chunks miss the surrounding context; longer chunks dilute the pattern signal. Aim for a self-contained moment.
Is this an LLM or pure pattern matching?
Pure pattern matching — no LLM. The tool runs locally in your browser, costs nothing, and produces consistent scores. The trade-off: it can't score nuance the way a model could, but it's 100× faster.
What's a "ship-it" composite score?
65 or higher across the 5 patterns. Below 40 is "weak" — likely won't retain. 40-64 is "decent" — worth A/B testing against a hook rewrite.
Does Clipflow use these same patterns to find clips automatically?
Yes. Clipflow's clip-finder scores every 30-90s window across the full episode using the same 5 patterns plus 3 additional ones (audio energy peaks, speaker transitions, sponsored-segment exclusion).
Related tools
Author + maintainer
Adrian Berisha — Founder of Clipflow. Indie SaaS builder shipping creator tools full-time since 2024. All tools and benchmarks on this page are reviewed quarterly. Last review: May 2026.
@clipflow on X · LinkedIn · hi@clipflow.to
Want this baked into your workflow?
Clipflow runs the whole repurposing pipeline — clip-finder, brand-voice captions, scheduler, multi-platform publish — for free on the Starter plan.
Try Clipflow free