Don't guess, just listen

Know your buyer.
Ship the ad today.

AlsterAds researches who actually buys your products, then writes, directs and renders the ads for them: realistic presenters, multi-shot video, statics and native, with the ad library and competitor tracking beside them. One platform, a fraction of the time.

From $59/mo · cancel anytime · every run priced on screen before it runs

“looks nice but will this snap in a month too, that's all I want to know”

2★ review · found by the research · example

becomes a persona

Persona · example

The one who got burned last time
trigger the old one brokeobjection durabilitywants proof at a year

the persona writes the ad

UGC · Talking head · 8s · example

this one opens on her exact words · every ad is written for someone specific

Made in AlsterAds

Every ad here knows who it's talking to.

UGC · Talking head
for: The one who got burned last timeopens on her exact words · example
UGC · Talking head
for: The upgrader who already owns twoangle: comparison · example
Product in use
for: The commuter on her third bagwritten to the proof she wants · example
Voiceover
for: The owner of a senior catnarrated in the persona's language · example
UGC · Talking head
for: The gift buyer with one weekendhook from his objection · example
Product in use
for: The owner who's tired of cleaning upopens on her exact words · example
Product in use
for: The owner of a senior catno presenter, product only · example
Cinematic
for: The owner who reads every 2★ reviewthree cuts, sound bed · example

Clips rendered in AlsterAds. Each was written for a researched persona. Some open on the buyer's exact words; all of them talk to someone specific.

The old way

The research happens last.
In the ad account, with your money.

a week

Guess the angle in a brief

Someone on the team decides what buyers care about. Nobody asked a buyer.

weeks

Book creators, ship product, wait

Then edit, caption and resize per placement, in four more tools.

after spend

Launch, then learn the angle was wrong

The first honest signal arrives as a CPA you paid for.

AlsterAds is the whole run in one place: research, script, direction, presenters, video, statics, native, with the ad library and competitor tracking beside them. A fraction of the time.

buyer surveysad spy toolcreator marketplacevideo editorcaption toolcopywriter→ one platform

How it works

Your brand's URL, once.
The research does the rest.

No brief, no shoot day, no handoff. AlsterAds pulls your brand, researches each product's buyers, and hands you ads that come filled out. Every step is priced in credits, shown on screen before it runs.

  1. 01

    Paste your brand's URL, once

    AlsterAds pulls your products, logo, colours and brand DNA. Every product is ready to research.

    brand set up once
  2. 02

    Run the research on a product

    Five to seven buyer groups. Threads, reviews and forums read with their comments. It finds every reason someone might want the product, including the ones you never thought of.

    6070 credits · priced before it runs
  3. 03

    Buyer groups come back, in depth

    Each persona with who they are, triggers, objections, the proof they want and the angles that fit. Not the obvious customer only.

    included in the run
  4. 04

    Everything comes filled out

    Script, shots, camera, voice and copy, chosen by the creative director for that persona. Change anything, or nothing. Then render.

    images from 3 credits · talking clips from 40
Example run · plays on its own · click a step to jump
yourstore.com
Products3 found · Everyday strap · Weekend bag · Card holderpulled
Logologo + coloursread
Brand DNAvoice, values, audience, competitorswritten
reddit · example“the last one lasted three weeks and then the clip just went”
review · 2★“looks nice but will this snap in a month too”
forum“is the strap actually leather or is that just the photo”
reply“my brother's is two years old and still fine”
q&a“does it fit a 20 mm lug, the listing doesn't say”

5 sources read with their comments · quotes kept verbatim · 5–7 buyer groups found · example

Persona 01

The one who got burned last time

will this snap in a month too
Persona 02

The gift buyer with one weekend

need it by Friday
Persona 03

The upgrader who already owns two

what does this do that mine doesn't

5 buyer groups returned · 2 you hadn't thought of · each with quotes, objections and angles

ScriptWill this snap in a month too? Here it is, one year in. Still fine.”auto
Shots3 · close-up, product in hand, wideauto
Camerafront, eye level → three-quarterauto
Voicewoman, 30s, warm · English USauto
rendering clip 1 of 3

What a research run returns

Your buyers already said it.
We find where.

Before a word is written, AlsterAds reads real discussions and reviews around products like yours. It finds every reason someone might want the product, including buyers you never thought of, and returns each group in depth: quoted, not paraphrased, with the angle it needs.

Reddit threads posts, comments, replies Forum discussions threads and replies Product reviews the two-star ones too Community Q&A questions before buying
Screen recording · a real persona research run
Screen recording of a persona research run returning buyer groups with verbatim quotes, objections and angles.

Quoted, not paraphrased

Pains, triggers and objections kept in the buyer's words, so the hook can be one they already said.

Buyers you never thought of

Every run starts from five to seven hypotheses, so it returns groups you would never have briefed: the gift buyer, the upgrader, the one replacing a failure.

A full profile, not a label

Who they are, pains, situation, objections, buying triggers, the language they use. Every builder reads it.

The output

Realistic presenters.
Cuts that hold together.

Super-realistic presenters with natural delivery. Multi-shot ads that keep the same person, face and energy across every cut. The best engine for each format is picked for you. It's all baked into the builders.

cut 1 · close-upCut one: the presenter in close-up, product in hand
cut 2 · three-quarterCut two: the same presenter from a three-quarter angle
cut 3 · wide, seatedCut three: the same presenter seated in a wide shot
One take, three cuts, one presenter.stills from one multi-shot render · example

Picked per format, automatically

Veo 3.1Talking presenters, the most natural voice we ship
talking head · podcast
Seedance 2.5Cinematic spots, three cuts with a sound bed
cinematic · multi-shot
Kling v3Product motion, with or without a presenter
product in use
GPT Image 2Statics with the headline composed into the artwork
static · native

You never pick an engine. The builder picks the right one for the format and your plan.

UGC · Talking head
Talking headone presenter, 8s · from 40 cr
for: the one who got burned last time · opens on her words
UGC · Multi-shot
Multi-shotthree cuts, same presenter · Growth and Agency
for: the upgrader · the side-by-side on cut two
Podcast
Podcasthost question, guest answer
for: the gift buyer · his objection becomes the question
Product in use
Product in useproduct motion, no presenter · from 8 cr
for: the owner of a senior cat · the proof she asked for
DualCast
DualCasttwo presenters, one conversation
for: the commuter · a friend asks what changed
Same presenter, every clip

Cast once, keep them. The same face and energy across scenes, angles and cuts.

Twelve languages and accents

US and UK English, four Spanish accents, German, French, Italian, two Portuguese, Dutch. Spoken, not dubbed.

Captions burned in, four styles

The persona's line on screen, styled to the placement.

Every placement

Talking clips in 9:16 or 16:9. Statics in 1:1, 4:5 and 9:16 from one batch.

Directed for you

The creative director fills it in.
You can change anything.

Script, shots, camera angles, movement, voice and copy come chosen for the persona. Edit nothing, or everything. The clip below changes with your choices.

PersonaThe one who got burned last timeauto
ScriptWill this snap in a month too? Here it is, one year in. Still fine.”auto
Shots3 · close-up, product in hand, wideauto
Copyheadline, primary text, three variantsauto
Captionsburned in · bold popauto

Or tweak it

Scene

Camera angle

Voice

Language and accent

and seven more

The direction the engine receives

priced before it runs · from 40 credits per 8-second clip · every clip keeps the same presenter

Home, morning · example

the clip swaps with the scene · example takes

Get started From $59/mo · cancel anytime

Static ads

The headline is in the artwork.

Batch image ads with the headline composed into the picture, not pasted on top. Every placement size from the same persona research, brand colours respected, product never recoloured. Runs on GPT Image 2.

Static ad: headline composed into the artwork
Static ad example
Static ad example
Static ad example
Static ad example
Static ad example
Static ad example
Static ad example
Static ad example
Static ad example
Static ad example
Static ad example
Static ad example
Static ad example
Headline composed into the image GPT Image 2Every placement size, one batch 1:1 · 4:5 · 9:16From 3 credits an image

Native ads and long-form copy

Native ads work.
Most brands haven't caught on.

An image that reads as an organic post, paired with long-form story copy people actually read. Scaling brands run native beside image and UGC ads. AlsterAds writes the story for the persona, in the buyer's researched language, and ships it with the image.

Native ad image: reads as an organic post
Native ad image example
Native ad image example
Native ad image example
The long-form pieceimage + story ship together · example

I wasn't going to post this, but my sister kept asking what changed. Six months ago I had honestly given up on fixing it. Then one comment buried in a thread sent me down a rabbit hole I did not expect to come out of.

Some context, because I think the context is the whole point. I wear the same watch every day. It was my grandfather's, it is not worth much, and I would not sell it for anything. The strap that came with it died a decade ago, and since then I have gone through, I counted last night, seven straps. Seven. Two of them were "genuine leather" from the market stall near my office, three were the kind you find at the bottom of a search page for nine dollars with four thousand reviews, one was a hand-me-down, and one I paid properly for and still lost within a year.

They all failed the same way. Not the leather, or not only the leather. The clip. That little spring bar and the clasp it snaps into. The strap would look fine, honestly it would look great, and then one morning I'd bend my wrist to pick up a coffee and feel the watch go loose. Sometimes it caught on my cuff. Twice it did not, and twice I spent twenty minutes on my hands and knees in a car park. If you have ever done that with something that was your grandfather's, you know the specific kind of sick feeling I mean.

So when I say I had given up, I mean I had decided that straps were a consumable. Like razor blades. Buy cheap, expect three weeks, keep a spare in the drawer. My sister thought this was insane, and she was right, but I could not see another option that did not involve spending more money on a thing I had already been burned by. The thing nobody tells you is that the last one lasted three weeks for a reason, and it was not me. I read every two-star review I could find before I ordered again, and the pattern was the same clasp, the same month, the same shrug from support.

That is what I actually want to talk about. The two-star reviews. If you only read the five-star reviews you learn nothing, because a strap is beautiful on day one. Everyone is happy on day one. The two-star reviews are the people who came back at week three, and I noticed that almost none of them complained about the look. They complained about the same two things: the clip, and what happened when they wrote to the company. Over and over. "Snapped in a month." "Support said normal wear." "Looks nice but will this snap in a month too, that's all I want to know." That last one was a review of a strap I nearly bought. I did not buy it.

The thread that changed things was not even about straps. It was a thread about a bag, of all things, and someone had written a long comment about how the failure point on anything you carry every day is never the material, it is the hardware, and that you should judge a product by its most boring component. Somebody replied with a photo of their strap, two years old, still fine, with a close-up of the clasp. Not a stock photo. A real, slightly blurry, kitchen-table photo with a mug in the corner. It turned out to be the same one my brother had been wearing for, he told me later, twenty-six months.

I want to be careful here because this is the part where these posts usually turn into an ad, and I promised myself it would not. So here are the boring specifics. The strap I ended up with is the everyday one, twenty millimetres, in the brown that is closer to tan than the pictures suggest. I paid full price, on a Tuesday, from my phone, and it arrived on the Thursday in a box that was, I will say it, nicer than it needed to be. The leather is stiffer than the cheap ones for the first week. That is not a complaint, it is just true, and if you like a strap to feel broken in on day one you will not like day one.

Before I ordered I did the thing I now do with everything: I wrote to support with a slightly stupid question. Not because I needed the answer, but to see how long it took and what the tone was. I asked whether the clasp could be replaced on its own. A person, with a name, replied in a few hours, said yes, told me how, and did not try to sell me anything. Compare that to the "normal wear" people. I am not saying support is the product. I am saying it is the second most boring component, and by then I had learned to judge things by their boring components.

What I actually cared about was the clip. The spring bars are the quick-release kind, the ones with the little tab, and they are thicker than any of the seven I had before. The clasp closes with a click you can hear across a quiet room. I know that sounds like nothing. It is not nothing. Every strap that failed on me had a clasp that closed with a sort of shrug.

I made a note in my phone the day it arrived, because I was fully expecting to write an angry follow-up at week three. Week three came and went. Then week six, which is when the second-cheapest one had died. Then the month mark, then two, then I stopped checking, which for me is the highest compliment I can pay an object.

Somewhere around month four I realised I had stopped thinking about it, which is a strange thing to notice about a strap. I had spent six years with a small background worry every time I lifted my arm, and it had simply gone. Nobody sells you that on a product page, and I am not sure they could.

So here is the one-year mark. Same one. Still fine.

I am not going to pretend it looks new. It doesn't. It has darkened around the holes and there is a lighter band where my cuff sits, and honestly I like it more now than I did in the box. The stitching is where it was. The clasp still clicks. I have not once had the watch go loose, not in a car park, not anywhere, and I have stopped carrying a spare in my bag, which I realise now I had been doing for six years without thinking about it.

A few things I would tell my six-months-ago self, in case you are him. One: stop reading five-star reviews. Read the two-star ones and look for the pattern. If the pattern is the hardware, walk away, no matter how good the leather looks. Two: if a company's response to "it snapped" is "normal wear", that is the review. That is the whole review. Three: a strap that costs four times as much and lasts twelve times as long is not expensive, it is just a different shape of the same money. I did the maths on the seven. It was not close.

Four, and this is the one I did not expect: the reason I finally believed it was not a brand, or a spec sheet, or a thirty-day guarantee. It was my brother's, two years old, on his wrist, at a barbecue, with a clasp I could see and pull on. That is the only reason I tried it in the first place, and it is why I am writing this now. Somebody's slightly blurry kitchen-table photo did more for me than every product page I had ever read.

A couple of honest caveats so nobody buys this on my word alone. If you have a very small wrist the tail is long and you will want the shorter size, which exists, I just did not know. If you want it soft on day one, it is not. And if you are the kind of person who swaps straps every week for fun, this is probably more strap than you need, and the quick release will matter more to you than the clasp does.

But if you are me, if you have a watch you would not sell and a drawer full of straps that failed the same way, and if you have honestly given up and decided that three weeks is just what a strap is, then no. It isn't. Here is the one-year mark, and I would have wanted to read this a year ago.

Edit: a few people asked which one. I have added it below. I do not get anything for this. My brother's is two years old now, which is the only reason I tried it in the first place, and I am writing this because I would have wanted to read it.

1,500+ words in the buyer's own language · written for the persona, from the same research · the image and the story come together

Also on every plan

And when you want to see
what is already working.

The market is one search away: live ads, the stores behind them, and the brands you track. On every plan, beside the research.

Ad discovery

See every ad your market runs, and which refuse to die.

Search live ads and filter hard: days running, format, niche, market. An ad still spending after ninety days is an ad that pays. Save it to a project, or clone the concept for your product.

  • Live ads, indexed and growing daily
  • Filter by run time, format and niche
  • Save winners to a project, then clone them
Screen recording · Ad discovery
Searching the ad library and filtering by format and run time.

Shop intelligence

See which stores are scaling.

Traffic estimates, live ad counts, and total ads for every store at a glance. Filter by platform or Shopify theme. Weekly Gems and Top Scaling surface the stores on the move.

  • Traffic estimates and live ad counts
  • Filter by platform or theme
  • Weekly Gems and Top Scaling
Screen recording · Shop intelligence
A shop profile with traffic estimate, live ad count and filters by platform and theme.

Competitor tracker

Their entire strategy, exposed.

Track any brand and get the whole picture: live ads and ads launched, charted over time. Their tech stack and top landing pages. Their best ads by longest running and most duplicated.

  • Live ads and launches, charted over time
  • Tech stack and landing pages
  • Best ads by longest running and most duplicated
Screen recording · Competitor tracker
The competitor tracker charting live ads and launches over time.

Research meets proof

Start from a winner,
in your buyer's words.

Pair the research with concepts that already work. Save a winner from the library and clone its concept for your product, spoken to your persona. Or double down on your own best ad before it fatigues.

Clone & Remix

Clone & Remix

Save any winner from the library and clone its concept for your product. New presenter, new product, your brand, your persona's language. The winning formula survives.

  • Pick a saved winner from the library
  • The concept is analysed, shot by shot
  • Rebuilt for your product and your persona
from 2 credits to analyse
Double Down

Double Down

Your best ad already proved what works. Spin it into fresh variations that amplify the exact concept that made it win, before it fatigues.

  • Pick your own best-performing ad
  • New hooks, scenes and presenters, same concept
  • Every placement size, ready to test
from 2 credits to analyse

Flows canvas

Ultimate freedom,
on an open canvas.

An open canvas of nodes: prompts, uploads, saved winners, personas, image and video generators. Wire them together and generate. Personas shape the on-image text and avatar scripts. Stitch multiple videos into one. Every canvas saves, and you can open as many as you need.

Screen recording · Flows canvas
The Flows canvas: nodes for prompts, personas, saved ads and generators wired together and run.

Personas shape image text and avatar scripts

Drag a persona onto a builder node and its words become the script and the on-image text.

Image, video, stitching

Render two takes, stitch them into one, add a voiceover, all on the board.

Edit and build on it anytime

Every canvas saves. Come back, swap a node, re-render only what changed. Open as many as you need.

What advertisers are saying

Built for people who ship ads for a living.

“The persona research pulled exact sentences off Reddit that I would never have written myself. We used one as a hook, word for word.”
Priya S.Head of growth · Skincare DTC
“I pasted a product URL between two calls. Before the next one I had a full set: image, video, copy, every placement size.”
Devon C.Ecom operator · Home & kitchen DTC
“It surfaced a buyer group we had never targeted in four years of running this brand, and nobody else is talking to them.”
Karim H.Founder · Supplement DTC

Pricing

One subscription.
Credits for what you make.

Research, tracking and the ad library are on every plan. Credits pay for generation and roll over, up to one month's allowance.

Starter
$59/month · USD
500 credits a month
For example1 research run, 60 static ads and 5 talking clips.
  • 25 tracked brands
  • 1 project, unlimited products
  • Single talking clips and product b-roll
Start on Starter
GrowthMost picked
$99/month · USD
1,500 credits a month
For example3 research runs, 150 static ads and 18 talking clips.
  • 100 tracked brands
  • 10 projects, unlimited products
  • Premium video: multi-shot, podcast, cinematic
Start on Growth
Agency
$249/month · USD
5,000 credits a month
For example10 research runs, 400 static ads and 70 talking clips.
  • 250 tracked brands
  • Unlimited projects
  • Premium video: multi-shot, podcast, cinematic
Start on Agency

Billed monthly in US dollars. Cancel anytime; downgrades take effect at the next renewal. Need more in a month? A 750-credit pack is $59, on any plan. Compare plans in full →

Questions

Fair questions,
straight answers.

Isn't this just another AI video generator?
The video is what comes out. What goes in is research: real threads, reviews and forum discussions about products like yours, read with their comments and quoted verbatim. Every script, cast decision and angle is built for a persona from that, not from a prompt you had to invent. And the video itself is built to hold up: realistic presenters, cuts that keep the same person, the right engine per format.
Do I have to write or direct anything?
No. Script, shots, camera angles, voice, copy and captions come filled out for the persona. You can change any of it, or render as is.
Where does the research come from?
Reddit threads with their comments and replies, forum discussions, product reviews and community Q&A. Each run starts from five to seven buyer hypotheses so it finds buyers beyond the obvious one. Quotes are kept as written, with the source.
What does a run actually cost?
Persona research is 6070 credits per run. Images start at 3 credits and talking clips at 40. The exact price is shown on the screen that carries the button, before anything runs. Unused credits roll over, up to one month's allowance.
Can I try it without a plan?
No. Every render costs real compute, so plans start at $59/mo with 500 credits and you can cancel anytime. Research, tracking and the ad library are included on every plan.
Can I cancel, or change plans?
Cancel any time from settings. You keep access to everything you generated, and unused plan credits are paused and come back when you reactivate. Upgrades apply immediately; downgrades take effect at the next renewal.

Start here

Your buyers already said it.
Go find where.

Paste your brand's URL, read your buyers in their own words, and have the first ad on screen this afternoon.

From $59/mo · cancel anytime

From $59/mo · cancel anytime Get started