Framer AI AgentsDesign and publish professional sites with AI
KREV
AI creative agents for ecommerce brands
416 followers
AI creative agents for ecommerce brands
416 followers
KREV helps ecommerce brands turn a single product image into product photos, video ads, and launch-ready creative. Unlike generic AI tools, KREV is guided by real ad signals, proven creative patterns, and tracked brand data to generate assets that feel more intentional, on-brand, and performance-ready.
Payment Required
Launch Team / Built With
KREV
Hey Product Hunt 👋, I’m Modi, CTO of KREV.
I built KREV because ecommerce brands are forced to grow across too many disconnected tools, and with KREV 1.0 we’re starting by solving the creative bottleneck.
There are plenty of AI tools that can generate images, but most of them create without enough direction. The result often feels generic, over-styled, or disconnected from what a real brand would actually run.
KREV takes a different approach.
Instead of generating from a blank canvas, KREV is guided by real ad signals, winning creative patterns, and tracked brand data. That gives the system a much stronger starting point and helps it produce creative that feels far more intentional.
💡 Why that matters:
• Static ads work when the structure works. Hook, hierarchy, layout, product framing, and offer presentation all matter. KREV uses real creative patterns to generate statics that feel much closer to actual performance ads.
• Video ads give brands more room to show the product, use case, and emotion behind it. KREV helps generate video creatives that feel more like campaign assets, not just motion for the sake of motion.
• Product photos are not just about making something look nice. Under the hood, KREV breaks down the product itself, studies placement, lighting, and composition, and generates imagery that feels more directed and premium.
🔮 Where we’re headed:
The long-term vision is a single workspace where ad performance data across Meta, TikTok, Google, and more feeds directly into creative generation and more. Your campaigns inform your creatives, and your creatives improve your campaigns. One loop, one place.
🙏 We’re still early, but the product is live and improving fast. I’d love your honest feedback on what stands out, what feels missing, and where you think we should take it next
Scade.pro
@modii Congratulations on the launch—it's a great product
KREV
@maria_anosova Thanks Maria!
The static ads from the Creative Agent are the best thing about KREV. Proper hierarchy, copy placement, layout structure — that's what separates you from every other AI image tool.
The problem I have with it is that: none of that is accessible via API.
POST /images gives me product photos. Great quality, but no headlines, no CTAs, no copy, no layout. The finished ads with text baked in only exist through the chat UI.
I'm building an automated creative pipeline for my e-commerce brand — agents generate briefs, produce ads, route approvals, push to Meta/Google/TikTok. KREV is the perfect rendering engine, but I can't have my agents chat through a UI.
What's missing:
POST /api/v1/ads
{
"product_id": "...",
"headline": "Headline here.",
"cta": "Shop Now",
"platform": "meta_feed",
"style": "minimal_dark"
}
→ Returns a finished static ad, same quality as the Creative Agent output.
That endpoint is the unlock. Every brand automating creative at volume needs this, not just the ones clicking through the UI.
KREV
@appelton Thanks for the feedback and it's a really good point. I'll surface the /ads endpoint into the API as well.
@modii Awesome thank you so much. I'll give it a try today and test everything. So far I really like the concept. Congrats on launch.
Launching on your birthday is a bold move, congrats Modi. One thing I'm curious about — when you say it's guided by real ad signals, does that mean it pulls from actual Meta/TikTok performance data or is it more like trained patterns baked into the model?
KREV
@abhra_das1 Thank you, indeed it is a bit nerve wrecking.
It’s a mix of both. We pull live data from a curated set of brands on Meta to understand what’s actually running right now, and then layer that with structured creative patterns we’ve built from high-performing ads.
So instead of generating from scratch, KREV starts from signals and proven frameworks which is why the output tends to feel much closer to actual ads.
I've seen too many ecommerce brands use AI to create product photos that look cool but completely bomb in ads because they weren't designed with performance in mind. Starting from a single product image and getting launch-ready creative that's actually informed by what converts is way more useful than pretty pictures. We're not ecommerce specifically but we need a lot of ad creative for our launch. Does KREV work for app products too, or is it mainly built around physical product photography?
KREV
@ben_gend Great point, that’s exactly the gap we’re focused on.
And yes, this works for app products too.
The way I'd approach this is the “product” becomes your UI. You can feed in key screens or flows, and KREV treats those as the core asset.
From there, you can generate ad creatives using software/tech patterns or even UGC-style formats built around your UI.
Brand DNA still applies as well, so everything stays consistent with your product’s look and feel.
Would actually love to see this used for a launch like yours and I'd love to get feedback!
Been burned by AI image tools before that just make everything look like a stock photo. The part about structured creative patterns is interesting. Is that something you built manually or trained from actual ad performance data?
KREV
@aaresvictor It’s actually a combination of both. We’ve built a structured layer of creative patterns manually, based on how real ads are put together, things like hooks, layout, product framing, and hierarchy.
yeah this is the real gap, a lot of AI stuff looks nice but feels disconnected from actual ads. are you feeding performance data back into the system or is it more pattern based for now? @KREV @modii
KREV
@aadhitya_muralidharan Right now it’s a mix. We pull real-world signals from a curated set of brands and combine that with structured creative patterns.
The goal is to move beyond just pattern-based generation into something that continuously learns from what’s actually converting.
@modii that sounds cool, amazing stuff! All the best :)
ad signal guidance is the thing that'll make or break this. AI product photos that look AI-generated are everywhere now - performance-based differentiation is the actual moat.
KREV
@mykola_kondratiuk Definitely agree. I’d add one more layer to that as well: performance-based "on-brand" creatives.
Performance signals lag brand signals — by the time data confirms on-brand, the creative is already stale.
KREV
@mykola_kondratiuk That's correct!
right, and chasing performance signals before brand health catches up is where most optimization decisions go sideways.