Ads
Campaigns that keep working after you close the tab
Wudtr writes the ads, scores them before a penny is spent, launches them on Meta and Google, then reviews what happened every single day against rules you can read. It pauses what is failing on its own. It never moves your budget without asking.
- Every budget change waits for your approval
- New campaigns are created paused, always
- Your own Meta and Google ad accounts
How it runs
From a sentence about your product to a live campaign
You describe what you are selling, give it a landing page or an article you have already written, set a daily budget and pick how much rope it gets. Everything after that is work it does and shows you.
- 1
Say what you want
Traffic, leads, sales or subscribers. Point it at a landing page, an article from your content library, or just describe the product in a paragraph.
- 2
It researches
It reads what you pointed it at and works out the angles, the audience and the keywords worth buying, rather than asking you to supply them.
- 3
It writes and scores
Headlines, body copy and images. Three models grade every draft before you see it, and the weak ones are rewritten or thrown away rather than run.
- 4
You approve the creative
Nothing launches until at least one ad has your approval. This gate does not move, whichever mode you are in.
- 5
It launches, paused
Campaigns, ad sets and ads are built in your own Meta and Google accounts and created paused. You decide when spending starts.
- 6
It reviews, every day
Cuts what is failing, keeps what is working, and writes down what it did and why. Budget moves arrive as a proposal.
The rules, in the open
The numbers it judges by
Most tools describe this as a black box. These are the actual thresholds, so you can decide whether you agree with them before you hand anything over, and check its work afterwards. Target cost per result is yours to set.
| When | It does this | On its own? |
|---|---|---|
| An ad costs 3× your target per result | Pauses it | Yes |
| An ad spends 5× your target with no result at all | Emergency stop, checked every four hours. Never before the ad is three days old, because a new campaign is still learning | Yes |
| An ad beats your target by more than 30% | Proposes 20% more budget | Asks first |
| A keyword spends 2× your target with nothing, after 20 clicks | Pauses it. Never more than three a day, and never one carrying 15% or more of your clicks, which gets flagged for you instead | Yes |
| A Google campaign reaches 15 conversions | Switches it from chasing clicks to chasing conversions | Yes |
| A region costs 1.5× the rest | Bids 40% lower there before excluding it | Yes |
| You reach 25 conversions on Meta | Builds a lookalike audience from the people who converted | Yes |
| One platform is beating the other by more than 30% | Proposes moving 15% of the budget across, never starving either below a fifth | Asks first |
| An ad is two weeks old and its results are sliding | Flags it as tired and offers a fresh angle | Yes |
| A campaign that was delivering goes quiet | Reverses its own recent cuts, once a day at most. The watchdog on its own watchdog | Yes |
Every decision is written down with what it saw and what it changed, so a week later you can read why a keyword died rather than guess. Nothing touches a live ad without leaving that record.
Two modes. In review, it proposes and you approve, and only the emergency stop and the kill switch act alone. In autonomous, the cuts above happen without you, and every budget move still waits. Review is the default, and a per-campaign kill switch stops everything at any moment.
The creative
Three models argue about your ad before you ever see it
Anyone can generate a hundred headlines. The useful part is throwing most of them away. Every draft is graded on hook strength, clarity, whether it speaks to a specific person, and whether it looks like anything else out there.
Copy for each platform's shape
Meta headlines, primary text and descriptions; Google responsive search headlines and descriptions. Written to each format rather than one blob pasted into both.
Images, at the price you choose
Four image engines from about five credits each to about forty, so a first test and a flagship campaign do not have to cost the same.
Up to eight audiences at once
A Meta campaign can test eight ad sets in parallel, each with its own slice of the budget and a floor so none of them is too small to learn anything.
Where it stops
What it will not do without you
An optimiser that can raise its own budget is a liability. These limits are in the product, not in a policy document.
It cannot move money
Raising a budget, cutting one, shifting spend between platforms, putting money behind an organic post: all four wait for a person, in both modes, with no setting to turn that off.
It cannot start spending
Campaigns are created paused on Meta and Google. Even the autonomous mode launches paused and tells you it is ready. Going live is a click you make.
It cannot run an ad you have not approved
A launch with no approved creative is refused outright rather than filled in with the best guess available.
One switch stops everything
The kill switch halts the optimiser and blocks any new launch for that campaign while it is on. It is checked on every pass rather than at the next daily review.
It leaves new ads alone
The emergency stop never fires on an ad under three days old. A conversion campaign shows nothing while the platform is still learning, and cutting it then is the classic way to waste the money already spent.
It changes one big thing at a time
Rebuilding a campaign around a more valuable goal happens at most once a fortnight and once a month, and only ever moves up the ladder. Everything else is small and reversible.
Knowing it worked
The sale has to find its way back to the ad
An optimiser is only as good as what it can measure. Because the signup, the trial and the purchase all happen inside Wudtr, it can send them back to the platform that earned them rather than guessing from a click.
Conversions sent back
Registrations, onboarding, trials and purchases go to Meta through the conversions API and to Google through its data manager, retried automatically if a send fails.
A ladder, not a single event
It starts by buying the cheap signal because there is enough of it to learn from, and climbs to the valuable one once that is arriving often enough to optimise against.
Reports for someone else
Paid performance and organic growth combine into one client-ready report, with an optional written summary, for when the person paying is not the person running it.
Conversion tracking needs your pixel and ad account connected. It is not automatic on day one, and the product says so rather than reporting zeros as if they were real.
Before you ask
What you need, and what this is not
Better to read this now than to find it out after connecting an account.
Your own ad accounts
You connect a Meta Business account and a Google Ads account you already have. We do not resell ad accounts or sit between you and the platform. Google needs an ordinary ad account, not a manager account.
Meta and Google, autonomously
Those are the two the daily optimiser runs. TikTok has campaign management and reporting in the same place, but it is manual: no autonomous cuts, no daily review.
Copy and images, not video
The creative it generates on its own is written copy and still images. Video ads are not part of the automatic loop today.
A daily budget, from a dollar
Set one before launch, and each Meta audience gets at least a dollar a day so it has room to learn. The ceiling is whatever you set; we do not impose one.
Credits for the writing
Generating and scoring creative spends AI credits from your plan, the same pool the rest of the platform draws on. The ad spend itself is between you and the platform.
Patience, at the start
It will not judge a creative before a thousand impressions or five dollars, and not a keyword before twenty clicks. Early numbers are noise and it treats them as noise.
Start it in review mode and watch it work
The honest way to try this is to let it research, write and score, then read what it proposes before approving anything. If you do not like its reasoning you have lost some credits and an afternoon, not a budget.