Report & teardown

A report that has actually watched your video

Scores are the headline. Underneath: the scene where attention drops, the line that should have been the hook, the frame the thumbnail should use, the moment the audio gets hard to hear — each with the fix.

Start free Get a free Hook Score

In every run

Included, always

Frame Timeline — every scene, what's in it, when it starts and ends; click to play it.
Transcript — full, timestamped, any language, right-to-left aware.
Topics Covered — as a list and as an interactive mind map with an offline HTML export.
Key Entities — people, places, brands, products, with first-seen times.

Premium modules · billed per run

Add what you need

Hook — /100 with clarity, curiosity, pattern interrupts, first-5s power and a retention prediction.
Pacing — attentional capture, information velocity, rhythm, a strong/weak timeline and attention-risk zones.
Thumbnail — salience, text clarity, composition, a mobile-size test and clickbait risk.
Audio Health — intelligibility, effort, level, consistency, voice vs background, hard-to-hear moments.
SEO & Competitive — keywords, titles, tags, positioning, gaps, recommendations.

Hook 68/100 · B · 2.7s to hook · Bold claim Clarity 8/10 · Curiosity 7/10 · Pattern interrupts 5/10 · First-5s power 8/10 Improvement: introduce a visual 'threat' element — the script says 'at risk' and 'money in play'; showing the stakes immediately would lift retention. Pacing 62/100 · Acceptable 00:49–03:10 weak — extended duration with a consistent visual style Attention risk zone 01:08–02:04 — no visual change or information-density spike Audio Health 93/100 · Excellent Other noise · subtle · [0:09] [0:19] [0:34] [0:49] [1:08]

Read it like an edit list

Find the lowest score, start there

A strong video with a weak hook is losing viewers in the first seconds; a strong hook with weak SEO is being watched but not found. Each module tells you which moment to change and why. Then switch to Editorial feedback to write the fix on that moment.

Independent AI-generated analysis. Not affiliated with or endorsed by YouTube. May contain errors.

Depth & models

Auto, Standard, Detailed or Deep

Deep reads fine text, OCR, code and slides — for software demos, lectures and anything with a screen in it — at roughly twice the cost of Standard. Models are Google's live Gemini list (Flash, Flash-Lite, Pro families). Every run shows the estimate before you commit.

Questions people ask

Straight answers

What is included in every analysis?

Frame Timeline (scene-by-scene with timestamps), the full timestamped transcript in any language, Topics Covered with a mind map, and Key Entities. These are included in every run.

What are the premium modules?

Content Analysis (hook effectiveness, pacing), Expert Feedback (thumbnail, strengths, improvements), SEO + Competitive, Audio Health, and the Channel report. They are billed per run and you only pay for the ones you add.

How is Audio Health measured?

Perceptually — speech clarity, listening effort, loudness balance and background noise as a listener experiences them, never in dB or LUFS — with each hard-to-hear moment flagged and the fix named.

What analysis depths are there?

Auto, Standard, Detailed and Deep. Deep reads fine on-screen text, OCR, code and slides and costs roughly twice Standard.