RouteMind-Ai Trip Planner

RouteMind — An AI travel planner that turns vibes, videos, and budgets into smart, real-world journeys.

Comment

Inspiration

Travel planning today is fragmented: inspiration lives on Instagram or YouTube, budgeting lives in spreadsheets, and logistics live in booking sites. We wanted to build a system that thinks like a human travel planner—starting from vibes, reasoning through constraints, and ending with a grounded, realistic itinerary. The launch of Gemini 3 inspired us to go beyond chatbots and build an autonomous, multimodal planning agent.

What it does

RouteMind is an AI-powered travel planning system that turns text, images, or videos into budget-aware, day-by-day journeys.

Key capabilities:

Autonomous multi-agent planning (budgeting, itinerary creation, constraint auditing, and optimization)

Vibe-based trip generation from uploaded photos or travel vlogs

Real-time grounding using Google Search for flights, hotels, and place verification

Visual timelines, maps, AI-generated travel postcards, packing lists, and PDF exports

The result is not a suggestion—but a verified, optimized travel plan.

Gemini Integration (Core of the App)

RouteMind is built entirely around the Gemini 3 model family, selecting the right model for each task:

gemini-3-pro-preview powers complex reasoning tasks like day-by-day itinerary sequencing, multimodal vibe analysis (images/video), and strict budget verification using Thinking Configs.

gemini-3-flash-preview handles fast, arithmetic-heavy tasks such as budget decomposition and low-latency conversational assistance.

Grounding via Google Search & Maps ensures locations, prices, and routes are real and verifiable.

Gemini is not an add-on—it is the planner, auditor, optimizer, and visualizer at the heart of RouteMind.

How we built it

Frontend: React 18, TypeScript, Vite

Styling: Tailwind CSS, Lucide React

AI Core: Google GenAI SDK (@google/genai)

Maps: Leaflet / React-Leaflet

Visualization: Recharts

Export: jsPDF

The architecture follows a logic-layer pattern, separating UI, AI orchestration, and grounding to ensure reliability and scalability.

Challenges we ran into

Preventing hallucinated prices and locations

Designing multimodal inputs without breaking UX

Making AI reasoning visible without overwhelming users

Accomplishments that we're proud of

A true Action-Era AI system, not a prompt wrapper

Multimodal planning from videos and images

Zero-waste inference using smart caching and agent orchestration

What we learned

Autonomous agents need constraints as much as creativity

Grounding builds trust more than perfect language

Good UX makes advanced AI feel simple

What's next for RouteMind-Ai Trip Planner

Group trip collaboration

Personalized long-term travel memory

Smarter carbon-aware routing

Mobile-first experience

Built With

Share this project:

Updates