How it works
Build your audience
A four-step wizard: pick countries, set the size, describe the segment. AI proposes the rules — demographics, segments, traits — and you edit them. Every member gets a name, a bio, and a face.
Test up to five versions
Compose your post — text and image — then clone it into variants. A/B or A/B/X, up to five versions side by side.
Read the verdict
Every member reacts to every version. You get engagement scores, a winner, individual reactions with reasons, and a plain-language commentary on what to fix.
What you get back
Not a black-box score — a report you can argue with.
- Attention
- 82
- Like intent
- 74
- Comment intent
- 51
- Share intent
- 63
- Viral potential
- 58
- Audience impact
- 70
Maya R.
“The first line made me stop. I'd send this to a friend.”
Tomasz K.
“Clear and honest. I'd like it, not comment.”
Ana S.
“Feels like a pitch. I scrolled past the third one.”
AI commentary
Version #2 wins on attention and share intent: the opening line names the problem before the product. Version #1 buries the hook under context; version #3 reads as an ad.
Six metrics
Attention, like intent, comment intent, share intent, viral potential, and audience impact — scored per member, per version.
A winner
The highest-scoring version, called out. Side-by-side comparison for the rest.
Member-level reactions
Every synthetic member reacts — like, love, support, celebrate, funny, wow, sad, or angry — and states a reason in plain words.
AI commentary
A short, readable summary of what worked, what didn't, and why.
What our customers say
Real teams testing real posts before they publish.
We usually pick the creative that feels right. This time the numbers pointed at a different version — and we trusted them.

Kasia, LEGO minifigures museum, Kraków
The A/B comparison saved us one extra round of opinions. We just published the version the panel preferred.

Paweł, Dolls & toy cars museum, Kraków
Why a synthetic audience works
Every member is generated to match your declared segment: country, age, sex, segments, traits — with an individual bio and portrait. Each one assesses each version independently and states a reason. That is the difference from asking a chatbot: not one averaged opinion, but a panel of them.
It is a directional, explainable signal — a focus group you can afford to run on every post. It is not a guarantee of metrics, and we won't pretend otherwise.
Pricing
Credits, pay as you go. No subscription.
- 100 credits = $1
- 500 credits free at signup
- A 100-member audience costs about $4.20 — once
- Each 100-member A/B run costs about $0.51
Your free credits cover your first audience and your first A/B test.