AI Personal Travel Planner
AI travel companion that generates complete, personalized itineraries from a single sentence—flights, hotels, restaurants, activities—in minutes, not hours.
Most people spend 10+ hours planning a single trip—Google Flights in one tab, TripAdvisor in another, a Reddit thread about the best neighborhood, a half-finished Google Doc that never quite becomes a real itinerary. The research itself becomes a second job. By the time a booking is made, decision fatigue has replaced anticipation, and half the best restaurants are already fully booked. The deeper issue is that every trip starts from zero. Someone who has taken forty trips over five years has strong, settled preferences—always direct flights, never chain hotels, prefers walking neighborhoods over resort zones—but no existing tool retains or uses those signals. Expedia optimizes for clicks. Google Trips aggregates confirmation emails. Neither one knows that you spent four days in Tokyo in 2023, so they'll happily resurface every ramen district guide they served you last time.
Built for Solo Founders, Developers, Side Hustlers.
Suggested stack: Next.js 15 + TypeScript, Claude API (claude-sonnet-4-6), Supabase, Amadeus / Skyscanner API, Booking.com Affiliate API, Stripe. Weekend scope: about 10 hours.
The Problem
Most people spend 10+ hours planning a single trip—Google Flights in one tab, TripAdvisor in another, a Reddit thread about the best neighborhood, a half-finished Google Doc that…
The Solution
An AI travel companion that works like a knowledgeable friend who remembers everything you've told them about travel. You describe the trip in plain language—"Food-focused long…
Market Research
$317.4 billion by 2029 — the global personalized travel and experiences market is projected to grow from $144.45B in 2024 at a sustained 17% CAGR, driven by demand for tailored…
Competitive Landscape
Wanderlog — Fast-growing trip planning app with strong collaborative features and real-time updates. Popular among group travelers. Personalization is rule-based rather than…
Business Model
Free ($0) — 1 trip/month, 3-day max itineraries, basic preference quiz, no booking integration
Recommended Tech Stack
Next.js 15 + TypeScript — App Router for SSR itinerary pages, server actions for AI calls, Edge runtime for low-latency preference lookups.
AI Prompts to Build This
Copy these build prompts into Claude, Cursor, or your AI coding tool. Create a free account to unlock the full research behind them.
1. Project Setup
Build the weekend MVP of "TripGenie": describe a trip in one sentence and get a complete day-by-day itinerary shaped by what the app already knows about you. Stack: Next.js (App Router), TypeScript, Tailwind, Supabase (Postgres, Row Level Security, Auth with Google sign-in), Claude (Anthropic API) for the parsing and the plan. Deploy on Vercel. Tables (Row Level Security on, each user reads only their own rows): - user_preferences(id, user_id, hotel_tier, routing_prefs, dietary, past_destinations jsonb) - trips(id, user_id, request_text, destination, start_date, end_date, party_size, budget_level, created_at) - itinerary_days(id, trip_id, day_index, summary) - activities(id, day_id, slot, name, description, address, estimated_cost_cents, booking_url) where slot is one of morning, afternoon, evening Screens: /login, /plan (the sentence box and the itinerary), /trips (saved trips). Env vars (names only): NEXT_PUBLIC_SUPABASE_URL, NEXT_PUBLIC_SUPABASE_ANON_KEY, SUPABASE_SERVICE_ROLE_KEY (server only), ANTHROPIC_API_KEY. API routes: POST /api/plan-trip, GET and PUT /api/preferences, GET and POST /api/trips. A small middleware protects the signed-in routes. Do not build: billing or plans, live flight search, hotel booking or affiliate links, the post-trip debrief, a second login method. Done when: npm run dev starts, you can sign in with Google, and the four tables exist with Row Level Security on.
2. Core Feature
Build the one feature that proves TripGenie: one sentence in, a full itinerary out. 1. POST /api/plan-trip takes a sentence such as "Food-focused long weekend in Lisbon with my partner, mid-range budget, we love walking and hate tourist traps". Parse the destination, dates, travel style, budget signals and party size. 2. Load the user's stored preferences: past destinations, hotel tier, flight routing preferences and dietary needs. 3. Call Claude with a structured prompt that combines the request and the preferences, and return JSON: an array of days, each with morning, afternoon and evening activities (name, description, address, estimated cost, booking URL if known), plus one flight suggestion and one hotel recommendation. 4. Save the trip, its days and its activities, and stream the result to the page as a day-by-day view with expandable activity cards. Rules: label flights and hotels as suggestions with no live prices. Never invent a booking URL: leave it empty when unknown. Empty state: a new user sees the sentence box with an example filled in. Done when: the Lisbon example returns a 3-day plan with three slots a day, a stored dietary preference is respected, and the saved trip reopens from /trips.
3. Landing Page
Build a one-page landing site for TripGenie, an AI travel companion. Hero: "Your AI travel companion that plans the whole trip, not just the flights." A live demo input pre-filled with "Food-focused long weekend in Lisbon with my partner, boutique hotels, we love walking". One button: Start planning your next trip, which joins the waitlist. Sections: a before and after ("Old way: 47 browser tabs, 6 hours. New way: one sentence, 2 minutes."), how it works in four steps (describe, review, refine and book, debrief), and three clearly marked placeholders for testimonials to be replaced with real ones. Waitlist: store the email in a waitlist table in Supabase. No other service. Style: warm off-white #fcfaf7, dark charcoal text, clean sans-serif, no hero illustration, just confident type and a subtle gradient. Done when: the page renders on a phone and a submitted email appears in the waitlist table.4. Branding Package
Use a design or image tool for this one. A coding agent cannot draw a logo. Brand for TripGenie: a clean wordmark with a minimal route-path or compass-needle icon that works as a 16px favicon. Colors: warm sand #F5EFE0, deep charcoal text #1A1A1A, terracotta #C4622D for buttons and highlights, and soft sky blue #A8C8E8 as the secondary. Type: Geist or Inter for the interface and a warmer display face (DM Serif Display or Playfair Display) for hero headings only. Icons: line icons at a 1.5px stroke, a 24px base grid and rounded end caps. Deliverables: wordmark, icon, a type spec with minimum text sizes, and one launch graphic. Done when: each deliverable is saved in one folder and the icon is readable at 16 px.