ReplayAutonomous app testing for teams who ship at AI-speed
Promoted
Decode by Entropik
Know what they feel. Decide what to build, ship, or sell.
100 followers
Know what they feel. Decide what to build, ship, or sell.
100 followers
Decode by Entropik is a human insights platform for UX, consumer, and market research. Run AI-moderated interviews, prototype and usability tests, concept and shopper studies, and predict ad performance before launch. Webcam eye tracking, facial and voice emotion AI reveal what people truly feel, not just what they say. Trusted by 150+ enterprise teams to turn research into revenue, fast.
This is the 2nd launch from Decode by Entropik. View more
AI Creative Insights by Decode
Launching today
Upload any ad, banner, OOH, or video creative and AI Creative Insights predicts how people will respond before your media budget goes live. Get attention heatmaps showing where eyes land first, second-by-second emotion curves, and prescriptive suggestions on what to change and why.
Benchmark against category norms, compare variants side by side, and use Synthetic Audience to see predictions persona by persona. Built by Entropik, the team behind Decode.
Free Options
Launch Team / Built With
Decode by Entropik
Hey Product Hunt 👋 Lava here, founder and CEO of Entropik. We're back!
When we launched Mira, our AI Moderator that runs user interviews on its own, you helped it finish #5. Your comments shaped a lot of what we built next. Thank you for that. 🙏
Today we're bringing you something our customers have been using for a while.
Picture this: three versions of an ad on screen, a launch on Friday, and a room full of smart people debating which one "feels right." Someone senior makes the call. The budget goes live. Two weeks later, the numbers tell you whether you guessed well.
Traditional pre-testing helps, but it takes weeks and a big research budget, so most creatives still ship on gut feel.
That's why we built AI Creative Insights. It's been one of our core products for a while, and today we're bringing it to Product Hunt. Mira helps teams understand what people say. This one predicts how people will react to your ads before they ever see them.
Upload any creative (video, banner, social, OOH) and within minutes you see:
Where eyes land first, and what gets missed
How people feel, second by second, and where they drop off
What to change, why, and the expected impact
How your creative stacks up against your category
Both products come from the team behind Decode, the emotion and attention research platform trusted by 150+ global brands.
🎁 For the Product Hunt community: use code PHCREATIVE to test 5 creatives and get your results instantly. Want early access to Synthetic Audience, to compare reactions persona by persona? Write to marketing@entropik.io with the subject "PH Early Access."
My ask: run it on an ad you're unsure about, then tell us here. Did the prediction match your gut? Where did it surprise you? If you used Mira last time, we'd love to hear how the two could work together for you.
I'll be here all day replying to every comment. Let's do this again! 🚀
Mira by Decode
Hi Product Hunt 👋 I'm Mridhu, heading Growth Marketing at Entropik. I ran our Mira launch here last time, so it's great to be back for round two.
As a marketer, this is the product I wish I'd had years ago. I've approved more ads on gut feel than I'd like to admit. With AI Creative Insights, I can see where people will look, where they'll lose interest, and what to change before we spend anything.
Two things I'd love to hear from fellow marketers:
- How do you decide between ad versions today?
- What would make you trust a prediction enough to act on it?
If you want to try it on your own ads, the code PHCREATIVE gets you 5 creatives for free. I'm here all day; ask me anything!
Mira by Decode
Hi all, I'm Sumit, VP of Machine Learning & Data Science at Entropik. Skeptics especially welcome in this thread. 🙂
The fair question is: "Can a model really know where people will look?"
Think of it like a weather forecast for attention. The model learns from real attention and emotion data, gathered from real people watching real ads. When you upload a new creative, it forecasts how an audience is likely to respond.
And like a forecast, it's a prediction, not a guarantee. It won't replace testing with real people for your biggest campaign of the year. What it does is give the other ninety-plus percent of your creatives, the ones that never get tested at all, some evidence before they go live.
Ask me anything: how it's trained, where it's strongest, and where it's not. I'd rather you hear the limits from me.
Kai
Super interesting, using Google/Meta's ad library to get an initial data baseline here? Curious to understand what this tool bases itself on to evaluate an ad :)
Mira by Decode
Hey everyone! I'm Naga, Product Manager on AI Creative Insights.
The question behind almost every product decision we make: how quickly can a marketer go from "which ad should we run?" to an answer they're confident in? That's why the results show up as a heatmap, an emotion curve and a short list of suggestions, rather than a long report.
Launch day is where we learn the most, so I'd really value your feedback:
- What was confusing the first time you used it?
- What's the one thing you'd want it to do next?
Every comment here goes straight into our roadmap discussions.
Mira by Decode
Hi Product Hunt! I'm Bharat, Head of Engineering at Entropik.
When Lava and I started Entropik, we made one bet: how people feel about something matters more than what they say about it. Most of what we've built since has been about measuring that feeling, properly and at scale.
AI Creative Insights is where that bet gets fast. What used to take a research study now takes minutes: upload an ad, and see where attention goes and how emotion moves, second by second.
Behind that simple upload is the hardest engineering we've done: making predictions quick enough to use in a real campaign, and consistent enough to trust across images and video.
Engineers and builders, I'd love your questions about how it works. Marketers, I'd love to know where it felt slow or clunky. That's the stuff we fix first.