See how your post performs before anyone sees it
Test your caption and media against a synthetic audience matching your target market. Get a real engagement score, read sample reactions, and rewrite with AI — all before you burn ad spend on a post that flops.
"New 2026 hybrid SUV. 52 MPG. Available this month."
Personas who stopped all mentioned gas savings. Move that to the first sentence.
From draft to rewrite in three steps
The whole loop fits into a coffee break.
Pick an industry, enter your geolocation and goal. We build a synthetic audience that matches.
Upload media and a caption, paste a Facebook or YouTube URL, or write from scratch. Personas react in real time.
Get an engagement score, sample reactions, and AI rewrite suggestions. Save the winning caption to your library.
Why teams use it every week
Catch weak hooks before they go live. Fix the caption, rerun, publish with confidence.
Paste any public Facebook or YouTube URL and see how your audience would react to it.
Each persona responds in their own voice, grounded in their occupation and interests.
Every simulation ends with actionable suggestions and a one-click rewrite button.
Bookmark simulations and save rewritten captions for reuse across campaigns.
Attached media is analyzed by a vision model so personas react to what they actually see.
Pre-built audiences for your niche
Eight industries come seeded with real-feeling persona archetypes. Unmapped niches fall back to on-the-fly generation.
Start free, scale when you're ready
- 3 simulations per month
- 10 personas per test
- All 8 seeded industries
- Sample reactions + engagement score
- 25 / 50 / 100 / 250 / 500 / 1,000 personas
- AI rewrite with suggestions
- Facebook / YouTube URL import
- Vision-model media analysis
- Saved captions library + bookmarks
Common questions
No. We use multi-agent orchestration where each persona has its own grounded profile (age, occupation, interests, voice, scroll habits) and reacts through a separate LLM call. Large simulations run 25 to 1,000 independent personas. The vision model reads your attached media once and shares the interpretation with every persona, so they react to what they actually see.
It is directional audience feedback, not statistical market research. Use it to catch obvious misses, test variations, and get actionable feedback on hooks and CTAs — not to predict exact engagement numbers.
Not in v1. Personas are generated from your industry, geolocation, and goal inputs, or pulled from our seeded archetype library. Real-audience seeding from connected Facebook and YouTube accounts is on the roadmap.
No problem. If your industry is not in our seeded library, the simulator falls back to on-the-fly persona generation using your targeting inputs. It works for any niche.
Yes. Paste any public Facebook or YouTube URL and the simulator will fetch the caption and media automatically, then run your audience against it.
Stop guessing what works
Get 3 free simulations when you sign up. No credit card required.
"Your morning just got better. ☕ 50% off all lattes this week only!"
