We help startups take off.
Video ads and paid creative, built with AI and tested on real budgets. First cuts in days, for founders in the year after a pre-seed or seed round.
A marketing studio for the year after the round.
We build marketing for startups with AI, so video and ads arrive in days instead of months, and in volume rather than one at a time.
Then we do the part most people skip. We put every version in front of real people on small budgets, read the numbers, kill what sinks and put everything behind what flies.
Marketing that gets you off the ground, then keeps climbing.
Video ads, paid creative and the testing under them. Built with AI, judged by people, measured on real spend.
Startups that closed a pre-seed or seed round in the last year. Founders with one idea rather than ten, who have been circling the same problem for years.
One letter off the Wright brothers. On 17 December 1903 they flew on under a thousand dollars, nine days after a government-funded machine went into a river.
A flight nobody sees is a rumour.
So here is one you can watch. Tap it, sound on.
Start with one ad. Keep what flies.
We study the ads already working for you, or for the category if you’re new, and build a ready-to-run video on that structure. You run it. If it works, we make the next ones.
A ready-to-run ad
Script, voice, footage, captions, sound. A 9:16 HD file you upload to your paid social as it is.
Days, not months
First cut in days. Changes in hours, because the people who make it are on the thread.
Dozens of versions
New hooks, lengths and offers on the same spine, so the test has something to pick from.
The makers
You talk to the people who write and cut the ad. Nobody in between.
Four steps, run on a loop.
One sentence a stranger can repeat. Everything else hangs off it, so we settle it before a single frame is made.
AI gives us dozens of video and ad variants in days instead of months. Real footage where it has to be real, generated where it doesn’t.
Every variant goes out on a small budget. The numbers pick the winner, not the meeting, and we kill the rest fast.
Everything goes behind what worked. Then the loop starts again next week, from a higher baseline.
Every loop starts from the last reading.
The people who make the work are the people you talk to.
Two brothers, one studio. No account layer, no handover to a junior after the pitch.
Everyone gets told no.
We don’t tell you why it won’t work. We tell you how to make it work.
Every company worth building looks unreasonable first.
There is a stretch after the round closes where your investors believe, your team believes, and nobody else has heard of you. That gap is the entire job.
You didn’t raise money to be realistic.
Realistic was the government project. It cost seventy thousand dollars and went into the Potomac nine days before the Wrights flew for a thousand.
Twelve seconds is not a failure. It’s a reading.
They measured it, changed something, and dragged the machine back up the rail. Four times before lunch.
Conviction is the only asset that isn’t on your cap table.
Nobody funded them and nobody believed them. The bicycle shop paid for everything, and they signed every drawing jointly.
A flight nobody sees is a rumour.
They put a camera on the sand before they flew. Belief travels on footage, not on decks.
Two men from a bicycle shop.
Two years before the world changed
They quit.
Season after season the gliders came up short. On the train home in 1901, Wilbur told his brother that man would not fly for a thousand years.
He was wrong by nine hundred and ninety-eight.
Same morning. Same machine. Same two men. Two years later Wilbur stayed up for thirty-nine minutes.
Nobody remembers a first attempt. They remember the curve it started. We are not hired for a launch. We are hired for the second year.
Drawn to scale. One pixel, one foot.
Handbills that were never printed.
The Wrights told almost nobody. No posters, no crowd, five witnesses on a freezing beach. This is what the most important morning of the century would have looked like if anyone had thought to advertise it.
Man Flies
Kill Devil Hills, N.C.
Thursday, Dec. 17, 1903
Four flights
before noon
Longest — fifty-nine seconds
No horse. No balloon. No wires.
Wright Cycle Co. requires
A Motor
under 200 lbs · 8 horsepower
Ten makers
have declined
Very well. We shall build it ourselves.
Six weeks · cast aluminium
Notice to the scientific public
The Tables
Are Wrong
Two hundred wing shapes tested
in a packing crate, with a fan
Measure it yourself.
852 Feet
in fifty-nine seconds
Against a twenty-one mile wind
On twelve horsepower
Total cost of the programme
Under $1,000
The public is invited to witness
Powered
Flight
Admission free
Bring a camera
A flight nobody sees is a rumour.
Wright Cycle Company
1127 West Third St. · Dayton, Ohio
Bicycles
Built & Repaired
Aeroplanes also
Enquire within.
Set in the style of a 1903 broadside. Every fact on them is real; the handbills are not.
SUCCESS FOUR FLIGHTS THURSDAY MORNING ALL AGAINST TWENTY ONE MILE WIND STARTED FROM LEVEL WITH ENGINE POWER ALONE AVERAGE SPEED THROUGH AIR THIRTY ONE MILES LONGEST 57 SECONDS INFORM PRESS HOME CHRISTMAS
Twenty-seven words. Two typos. A request to be home for the holiday.
It didn’t need a landing page. It needed someone to see it.
Nothing in that shed was bought off a shelf.
If the tool doesn’t exist yet, that’s not a blocker. That’s the job.
The wind tunnel
A packing crate and a fan. Two hundred wing shapes tested, because the tables the whole world trusted were wrong.
Now: dozens of versions of your ad, tested the same way.
The warping cradle
Wires from the pilot’s hips to the wing tips. Control was designed first, not bolted on last.
Now: positioning before spend. We fix the message, then buy attention.
The propeller
Nobody could tell them how. So they reasoned it out: a propeller is a wing that travels in a spiral.
Now: video. What turns the money you already raised into forward motion.
The engine
Twelve horsepower, cast in aluminium, built in six weeks by the shop mechanic after every manufacturer refused.
Now: if the tool you need doesn’t exist, we build it rather than wait.
The stopwatch
Every glide logged. They picked Kitty Hawk off Weather Bureau records, not instinct.
Now: one number a week that decides what changes.
The camera
Aimed at the end of the rail and handed to a coastguardsman who had never used one. That frame is why the world believed.
Now: proof. Your real product, on camera, before you fly.








Neither of them finished school.
No degree, no patron, no laboratory. A workshop, a library card, an argument that ran four years, and each other.
What we take from them is not the nostalgia. It is the method: suspect the data, build the instrument nobody sells you, and put the camera up before you fly.
For some years I have been afflicted with the belief that flight is possible to man.
Wilbur Wright to the Smithsonian, 30 May 1899Wilbur twists an empty box in the shop and invents the way aircraft still turn.
They almost give up, then build a wind tunnel and prove the whole field wrong.
17 December. Four flights. Total cost: under a thousand dollars.
Le Mans. The people who called them liars print apologies.
Armstrong carries a scrap of the Flyer’s muslin wing to the Sea of Tranquillity.


Onboard a Wright machine · Italy · 1909
Tell us where you’re trying to get to.
Closed a pre-seed or seed round in the last year? Write two lines: what you make, and the number you need to move. We’ll answer with the first ad we’d build.
We reply to everything, usually the same day.