About this project

Built on shoulders of giants

An AI-native travel planner built to explore what happens when 8 specialized agents replace a search form — and what the industry leaders that came before us got right.

Where we got inspired

What Mindtrip + Sabre + PayPal got right

In a June 2026 webinar, Mindtrip demonstrated an end-to-end agentic flight booking flow — chat-first search, transparent reasoning, and a checkout backed by Sabre (shopping/pricing) and PayPal (identity + payment). Four patterns from that flow shaped how we built this:

  • 💬

    Chat-first entry, no rigid form

    A single natural-language input replaces multi-field search forms — reflected in the home page search bar and the concierge chat.

  • 🔎

    Transparent agent reasoning

    Live trace of which agent is working, on what, and whether it's done — visible in the agent trace panel during planning.

  • 🧭

    Narrative results, not just a grid

    Results come with a written rationale alongside structured comparisons — not just a table of prices.

  • 🗂️

    Trip as the central object

    Saved trips, price watches and itineraries are organized around a persistent trip record rather than one-off searches.

Conversational trip planning

Also inspired by

The tools and players we learned from

Google Flights

Price transparency and date-flexibility as core UX — not buried in filters.

Kayak Explore

Destination-agnostic search: 'show me anywhere warm for under €1,500' as a first-class query.

Airbnb

Emotional destination cards with photography-led UI instead of data-dense tables.

GetYourGuide

Activities as a first-class travel product — not an afterthought appended to flights/hotels.

Rome2rio

Multi-modal route transparency: show all options (fly, train, bus, drive) side by side.

Sherpa°

Visa and entry requirements as structured, always-verified data — not scraped or hallucinated.

Where we sit

Gen AI → Agentic AI → Agentic commerce

Travel AI is moving from recommendations to execution. Here's where this project sits on that continuum.

Stage 1

Generative AI

AI helps research, plan and decide — surfacing recommendations, drafting itineraries, answering questions. The traveler still books manually.

Fully implemented
Stage 2

Agentic AI

Agents execute multi-step tasks autonomously: searching flights and hotels in parallel, optimizing budgets, replanning when something changes.

Implemented (8 agents)
Stage 3

Agentic commerce

Agents research, negotiate and complete transactions — booking and paying — on the traveler's behalf, via providers like Sabre and PayPal.

Out of scope (by design)

Under the hood

8 agents. One perfect trip.

✈️

Flight Agent

Best flights, cheap stopovers, optimal times.

🏨

Hotel Agent

Prime location, best price-performance ratio.

🗺️

Activity Agent

Culture, adventure, restaurants, day planning.

💰

Budget Agent

Optimize total costs, find alternatives.

🔍

Research Agent

Visa, climate, safety, local tips — all before booking.

🚨

Disruption Agent

Flight disrupted? Status, alternatives and rights in seconds.

🎒

Packing Agent

Destination-specific packing list tailored to climate and trip type.

📋

Briefing Agent

Full country briefing: culture, currency, health, safety.

Try it yourself

From idea to plan in minutes.