Prompty is Duolingo for AI. Five-minute challenges, scored feedback on real prompts, XP, streaks, and a mascot that grows — turning a tool people own into a skill they use.
See the roundEvery major platform now ships an AI assistant. The bottleneck moved from access to ability — and the current answers to that gap don't stick.
A 4-hour prompt engineering course converts a tiny slice of intent into completion. Most people never finish week one.
Copy-paste packs solve one task and build zero skill. Users return to the blank box just as stuck.
Companies buy seats and run a one-time workshop. Usage flatlines by week three and the licenses churn.
Duolingo didn't win by being the best language pedagogy — it won by being the thing you open at 11:58pm to save a streak. Prompty applies the same loop to the most valuable general-purpose skill of this decade, in a category where the "practice partner" is the AI itself. That makes the feedback instant, personalized, and nearly free to deliver.
Users get a real task, write their own prompt, and receive a scored critique plus the stronger version. XP, badges, streaks, and a mascot that evolves at day 7, 30, and 100 carry them to tomorrow.
Open the product siteThe mascot is the text cursor itself — the thing users used to stare at. Progression is visible, personal, and costly to abandon.
Marginal cost per lesson approaches zero, so the library scales with usage instead of headcount. Human review sets quality bars per track.
Every attempt is a labeled before/after. Over time that dataset tunes grading, personalizes difficulty, and is not easily replicated.
Consumer freemium drives volume and word of mouth. Teams and institutions pay for the same loop with reporting attached — the piece every employer buying AI seats currently lacks.
One challenge a day, streaks, XP, mascot to Spark. The acquisition and habit layer.
Unlimited challenges, all skill tracks, prompt history review, streak repair, advanced grading.
Cohort dashboards, assigned tracks, completion reporting. Sold to employers and institutions already paying for AI licenses they can't measure.
Three curves crossed at once. First mover on the daily habit takes the category name.
AI is bundled into phones, browsers, and office suites. The addressable user is now everyone, not early adopters.
Model quality is finally high enough to critique a prompt usefully at a cost per session measured in fractions of a cent.
Budget holders are asking what their AI spend produced. That question has no product answer today.
Replace the placeholder figures below with current numbers before sending this to anyone.
Prototype live, waitlist at [X] signups, [X] challenges written across [X] tracks.
Ship iOS and web beta to [X] users. Target metric: D7 retention and average streak length, benchmarked against consumer learning norms.
Launch Prompty Plus, test price points, and run [X] paid team pilots to validate the B2B wedge.
Convert pilots into contracts with employers and higher-ed programs, and open the seed round on retention plus revenue.
Courses own intent, not attention. Prompt libraries own tasks, not skill. Model makers own the tool and have no incentive to build a third-party habit layer across all of them.
A 90-day streak is switching cost that no feature clone recovers.
Prompty teaches skills that transfer across every assistant, so platform shifts help rather than threaten.
Category-defining mascots compound. Duolingo's owl is worth more than its curriculum.
The single question that matters at this stage is whether people come back on day seven. This round funds getting to a clean answer.
Dashed figures throughout this page are placeholders. Swap in real numbers, or delete the metric, before sharing.