For YouTube Shorts creators

Five ready-to-shoot scripts a week. With receipts.

The hard part is not filming. It is knowing what to say in the first three seconds.

Every week: five Shorts scripts — the subject, the first three seconds written out word for word, the body, the call to action and the shot list. Each one is built on an opening that measurably recurred across independent channels, and every claim links the source videos it came from, so you can check us before you trust us.

Get early access What we measure

No channel connection. No filming for us. No account access — everything is measured from public data.

2niches measured
continuously

0Shorts tracked
since launch

0growth measurements
taken

0topics qualified
right now

weekly report · ai & tech news

Pattern · Cost before tool script 1 of 5 in this report

Length 15–30 seconds

“You are paying three AI tools to do one job.”

Then say

Name the overlap: writing, summarising and search are one workflow, split across three subscriptions.
Show one tool doing all three, on screen, in under ten seconds.
End on the monthly number that stops leaving your account.

CTA — “Comment which two you would cancel. I’ll test the winner next week.”

15source videos cited in this report

11independent channels behind them

3.4×best source vs its channel's normal

20%most any one channel contributes

The videos this opening is derived from

Video	Channel	Views	vs normal	Age
I cancelled my $20/mo AI subscription	Channel A	412K	3.4×	6 d
Stop paying twice for the same AI	Channel B	188K	2.8×	11 d
Nobody needs the paid tier for this	Channel C	96K	2.9×	3 d

Shot list

Screen recording: the three dashboards side by side.
Cursor hovering the cancel button on one of them.
Phone calculator with the monthly total, held to camera.

Illustration, not a screenshot. The lines and numbers on this card are invented to show the format. Real reports cite real videos, and every number in them is computed from the cited sources — never typed in by a model, never typed in by us.

The advantage

Other tools hand you an outlier. We hand you the next video.

Outlier finders — and there are good ones — show you a Short that beat its channel's average. True, useful, and still three decisions away from a video: what subject, which opening, what to say. We finish those three.

Where an outlier tool stops

A list of videos that overperformed, ranked by a multiple. The research is done — and the writing has not begun. You still face a blank page, now with a number next to it.

A multiple is not a video. Somebody still has to decide what to say.

Where your report starts

Five scripts built from that same measurement: the opening written out, the body, the call to action, the shots. Each opening follows a structure that recurred across independent channels — one channel's habit is a habit; several channels arriving at the same opening is worth copying. Every claim links its source videos, so checking us takes one click, not an evening.

Read, pick, shoot. The research and the writing are already done.

Behind the reports, the board watches topics the same way: a topic qualifies only once 3 independent channels reach it, with at least 2 of them beating their own normal, measured repeatedly as the group grows. Right now nothing clears that bar, the counter above says zero, and no candidate gets dressed up to fill the gap.

How it works

Four steps, no guesswork.

Pick one of 2 niches

AI and technology news, or Personal finance and side hustles. We maintain the vocabulary and the creator panel behind each one, so measurement starts the moment you sign in — there is nothing to configure.
We measure Shorts as they grow
Checked again at 6, 12, 24 hours and beyond. Most tools read a view count once; we keep the curve, so every video is compared against its own channel's normal result at the same age — arithmetic on measured data, replayable from its stored inputs.
Patterns need witnesses
From the last 90 days of measured Shorts we extract opening structures that recur across independent channels. A structure seen from one channel is discarded, and no single channel is allowed to dominate the evidence behind a report.
Five scripts, checked
A language model writes the script text — and nothing else. Every number is computed by code from the cited videos, every script is checked against its evidence before it reaches you, and a script that cannot cite its sources is dropped rather than published.

Straight answers

What this does not do.

You have been sold to by a dozen of these. Here is where ours stops, before you ask.

It does not promise views.We hand you evidence and finished writing. What you shoot with it is yours.
It does not call trends.No “rising”, no “early”, no “post this today”. The day a subject carries a timing claim, it will carry the measurement that licenses it — and that measurement is not done.
It does not touch your channel.No connection, no OAuth, no permissions. Public data only.
It does not invent evidence.A report that falls below its own bar — too few sources, too few channels, one channel dominating — refuses to publish. The bar stays; the issue waits.

Pricing

Early access is free. Pro opens the rest of the scripts.

Early access stays free: the board, both niches, every number linked to its source videos, and the first ready script in each weekly report. Pro opens the remaining scripts in that report. Nothing is written to order and nothing is produced on demand — these are the same measured reports, in full.

Early access

Free

no card, no expiry date

Both niches as they fill
The first ready script in each weekly report
Every claim links its source videos

Get early access

Monthly

$29 / €27

per month, cancel any time

Every weekly report in full — all five scripts
The evidence behind every script, source videos included
Cancel from your account page

Create an account

Yearly

$290 / €269

per year — two months not paid for

Everything in monthly
Priced for the long game