JoyRide.

The problem

Family trip planning is a fragmented, multi-hour slog crammed into 45 minutes of evening downtime. Parents Google "[destination] things to do with kids," open 6–10 browser tabs, cross-reference TripAdvisor, parenting blogs, Instagram saves, and try to stitch it all into a coherent plan — across 8–15 hours over 2–3 weeks per trip.

Then comes the emotional kicker: after all that effort, one kid takes one look at the finished itinerary and says "yeah, not interested." The plan is also scattered across tabs, screenshots, and mental notes — there's no clean, shareable output for the other parent or the family group chat.

The solution

JoyRide collapses the entire planning loop into a single form fill. Parents enter destination, travel dates, and children's ages (gender optional); the AI returns a complete itinerary clustered geographically, sequenced for the kids' energy levels, and matched to the weather forecast — with TripAdvisor badge integration for trust, address and description on every activity, and zero invented venues.

The killer feature for parents: the itinerary is shareable, modifiable, and trustworthy in one document instead of ten tabs.

How it works

Flow: User enters destination, travel dates, children's ages, genders (optional) → form input prevents prompt-injection edge cases and improves model behaviour → AI synthesizes itinerary with strict rule enforcement → user reviews, can swap activities via on-screen buttons (no dialogue navigation, less ambiguity).

Two-model Claude strategy: Claude Opus 4.6 for main itinerary generation (planning depth matters), Claude Sonnet for other activities and stores sections (faster, cheaper, quality sufficient). Tried Haiku first for cost efficiency, escalated when output quality required it.

Hard system prompt rules: no invented venues, no restaurants/hotels invented, no adult venues, no specific times (only suggested durations), no duplicate activities, age-appropriateness flags, stroller-friendly tags. Real family itineraries (provided as separate docs) are used as test cases, with edge-case coverage for weather, family composition, geography routing, partial travel days, and API failures.

Who it's for

Target user: a busy, tech-savvy parent (mid-30s to mid-40s) juggling a demanding career and multiple kids spanning a wide age range. The family takes 1–3 vacations per year (mostly road trips, occasionally fly-in trips). Currently spending 8–15 hours per trip across 4–8 sources trying to piece together an age-appropriate plan.

Use cases: city break planning, road-trip multi-stop itineraries, fly-in destination weeks. The product is most valuable for the mid-30s to mid-40s parent demographic with multi-kid families — the segment most acutely caught between career time pressure and the desire to make family travel actually fun.

Why it matters

Family travel is at a structural inflection. 92% of parents plan to travel with their kids in the next year — the highest level since the pandemic — and 100M+ Americans are expected to take family vacations. Meanwhile AI in tourism is growing at 28.7% CAGR through 2030, with the family-specific wedge pushed higher by booming demand and "kidfluence" purchasing decisions.

JoyRide's bet is that no incumbent travel app is purpose-built for the family-specific constraints — age-appropriateness, geographic clustering, weather logic, and verified venues — and that the combination of those four constraints is the moat. Generic AI itinerary tools will keep losing to family-specific ones because the rules are the differentiator.

At a glance

Project

JoyRide

Built by

Joy Mimun

One-liner

AI family travel planner — age-aware, weather-informed, geographically clustered, no-invented-venues itineraries for busy parents juggling multiple kids.