Vibe Travel (Built w/ Claude Code)

Family trip planning is broken. Parents spend hours combing through generic "top 10" lists that don't account for nap schedules, sensory needs, or the reality of traveling with a three-year-old. Vibe Travel fixes that — an AI-powered planner that learns your family's profile and builds real, personalized itineraries for destinations worldwide.

Context

The core insight was simple: every family is different, but every travel app treats them the same. Vibe Travel starts with a family profile — kids’ ages, travel pace, sensory needs, dietary restrictions, budget — and uses that context for everything. Search results, attraction recommendations, and day-by-day itineraries are all filtered through the lens of your specific family, not a generic traveler.

The app includes Scout, a conversational AI assistant powered by Claude, that can plan entire trips from a single prompt, answer questions about destinations, save attractions, and refresh itineraries in real time.

 

Project Info
Date:

2025-2026

Role:

Sole Builder & Tinkerer

Technologies & Tools:

Claude Code, Next.js, TypeScript, Tailwind CSS, Supabase, Anthropic Claude API, Google Places API, OpenWeather API, Viator API, Wikipedia API, Vercel

Activities and Responsibility:

  • Defined the full product — problem framing, user flows, feature prioritization, and build sequencing
  • Built the family profile system (“Vibe”) as the personalization engine across all features
  • Integrated 6 external APIs to enrich AI recommendations with real data (ratings, hours, accessibility, weather, tours)
  • Designed and built Scout — an agentic AI assistant with tool use that can plan trips end-to-end from a single prompt
  • Shipped full auth, database, and trip management system on Supabase
  • Deployed to production on Vercel

Key Features:

  • Family Vibe Profile — kids’ ages, sensory needs, pace, dietary restrictions set once and used everywhere
  • AI Attraction Search — 12–16 personalized recommendations per destination, enriched with Google Places data
  • Nap-time Aware Itinerary Generation — day-by-day plans that account for energy levels, meals, and realistic family pacing
  • Scout AI Assistant — conversational, tool-enabled Claude agent that can search, save, and plan in real time
  • Sensory-Friendly Filters — accessibility and sensory notes on every attraction
  • Live data from Google Places, OpenWeather, Viator, and Wikipedia

Outcome:

  • Fully functional, production-deployed SaaS at vibe-travel-six.vercel.app
  • Covers 50+ cities worldwide
  • Free to use with no credit card required
  • Demonstrates full product lifecycle — problem definition through shipped, live product

Build Process:

Discovery
  • Identified the gap between generic travel recommendations and the real complexity of family travel
  • Mapped user journeys for parents across the full planning lifecycle — inspiration to day-of execution
Build & Iteration
  • Started with the family profile system as the foundation — personalization only works if the context is right
  • Built search and attraction enrichment before itinerary generation, validating the data quality first
  • Added Scout as the agentic layer to tie all features together through conversation
  • Refined itinerary generation logic based on real family travel constraints
  • Expanded API integrations to reduce AI hallucination with grounded, real-world data