vacation planner(Elite Concierge)

Plan luxury trips in 30s. Elite Concierge uses Gemini AI to turn vacation dreams into professional dossiers with curated stays and validated budgets. From search to delivery instantly.

Comment

Inspiration

Planning a vacation often turns a moment of excitement into a "second job." We noticed that high-end travelers spend upwards of 10 hours toggling between flight aggregators, hotel reviews, and blog posts. We were inspired to build a tool that moves the needle from information retrieval to intelligent orchestration, creating a "one-click" experience that mirrors the service of a human luxury concierge.

What it does

Elite Concierge is an agentic travel architect. A client simply inputs a natural language description of their dream vacation. In seconds, the application synthesizes a comprehensive travel dossier including:

Curated architectural and cultural landmarks.

Luxury accommodation recommendations.

Real-time weather insights.

A programmatically validated financial breakdown with a calculated Grand Total for the entire journey.

How we built it

The project is powered by a Flask backend and a Gemini 3 Flash intelligence layer. We utilized several advanced AI patterns:

Single-Pass Synthesis: To optimize speed and API quota, we designed a "Super-Prompt" that forces Gemini to reason across multiple domains (culture, logistics, finance) simultaneously.

Structured Output Parsing: We used Gemini’s ability to follow strict schemas, generating pipe-delimited strings (Item∣Cost) that our Python engine converts into data objects.

ReportLab PDF Engine: We built a custom document generator that transforms AI "thoughts" into professional, high-fidelity PDF brochures.

Challenges we ran into

The biggest hurdle was the 429 Resource Exhausted error. Our initial "Swarm" architecture triggered five parallel requests per user, which overwhelmed the Free Tier API limits. We solved this by implementing Rate Limiting and consolidating our agents.

Additionally, we faced a "Silent Failure" with the PDF export. The browser was trying to read binary PDF data as text, leading to undefined errors. We resolved this by re-engineering the frontend to handle the response as a Blob (Binary Large Object).

Accomplishments that we're proud of

We are particularly proud of our Total Cost Calculator. It uses Regular Expressions to sanitize AI-generated text, ensuring that even if the AI writes "$1,200 USD," our system correctly extracts the float 1200.00 to perform the summation logic: Grand Total=i=1∑n​Costi​

This ensures the financial summary is always mathematically accurate.

What we learned

We learned that Prompt Engineering is as much about constraint as it is about creativity. By strictly defining the output schema, we reduced backend processing errors by over 80%. We also gained deep experience in managing asynchronous event loops in Python and handling binary streams in modern web browsers.

What's next for vacation planner(Elite Concierge)

The next evolution of Elite Concierge involves:

Multimodal Inputs: Allowing users to upload a photo of a place they saw on Instagram to trigger the itinerary generation.

Live API Integration: Connecting to flight and hotel booking APIs for real-time pricing.

Currency Localization Agent: Using IP-based geolocation to automatically convert the budget into the user's local currency.

Built With

  • flask
  • google-genai
  • python-dotenv
  • reportlab
Share this project:

Updates