Paste a public YouTube link. The analyzer watches the video, reads the transcript, and ties every observation to a timestamp — long-form gets judged by long-form rules, Shorts by Shorts rules.
A regular video or a Short. Public links only — no channel connection, no upload.
Frames are sampled across the whole runtime and the audio is transcribed, so longer videos take longer. A Short finishes in a couple of minutes.
Findings point at moments — the promise in the opening, the stretch where attention likely drops, the payoff. Each one is something you can recut.
Everything comes from the video itself — the frames and the spoken words — not from private analytics.
Long-form has fifteen to thirty seconds to earn the click's trust. The analyzer reads whether those seconds restate the title's promise or stall before it.
Where the video coasts — long unbroken takes, repeated points, sections that could be one sentence. Each flagged with a timestamp.
Whether the video actually delivers what the title and thumbnail promised. A mismatch there reads as a content problem even when the packaging did its job.
How the video ends: a reason to watch the next one, or a fade-out that wastes the viewers you already earned.
Extensions annotate the page you're browsing — useful for scanning metadata at scale, but they don't watch the video. This analyzer reads the actual content: what's on screen, what's said, and when. Nothing to install, works in any browser.
Yes. Paste a competitor's video or an outlier in your niche and read why it holds attention. Use it as a teardown: borrow the mechanism — how the promise is set, how sections hand off to each other — never the wording.
Split the question in two. If impressions are the problem, that's packaging and discovery — title, thumbnail, and topic choice; the analyzer can't see your impressions, but the title generator and thumbnail preview next door work that side. If people click and leave, that's the content side — and that's exactly what this report reads.
Yes, and it switches rules: a Short gets judged on a roughly three-second hook window and swipe-speed pacing, while a regular video gets the long-form treatment. Same link box for both.
This analyzer reads one video at a time. Kabo's Radar preview shows the intended next step — breaking down your niche's outliers into topics and scripts — but channel tracking and weekly direction lists are not live today.
Yes — three analyses a day with no account, shared across Kabo's video analyzers, entirely online with nothing to install, and the diagnosis, timestamps, and scores are free in full. A free account removes the daily limit, unlocks the action list (edits, alternative openings, title ideas), and saves every report to your history.
No. Those live in your YouTube Studio and stay private. The analyzer reads only the public video — which is exactly why it can also read videos that aren't yours.
No. A public link is the only input.
A few minutes for a Short, longer for long-form — the video is sampled second by second and transcribed before anything is judged. Videos up to 30 minutes are supported; anything longer is declined upfront rather than half-analyzed.
Private videos can't be fetched. Public videos always work; unlisted ones sometimes do, but public is the reliable path.
The report includes a summary of what the video does, but if a clean text summary is all you need, the video summarizer tool is faster and doesn't use an analysis run.
It won't pretend to read the algorithm's mind — nobody outside YouTube can. It shows you what a viewer saw at each moment and where that likely cost you watch time, which is the part you can actually change.
