Portfolio case study
Luna Blu Retreat is an end-to-end demo for turning stay requests into tracked opportunities: availability checks, estimated quote, AI summary, HubSpot CRM, email confirmation and operational logs.
Real problem
Hotels, B&Bs, rentals and local services often handle requests from websites, email, phone and WhatsApp. The team must check dates, guests, extras, price and sales priority before it can even reply. The demo shows how to turn that repetitive work into a measurable operating flow.
Solution
The important part is not only the form. The frontend calculates availability and quote with explicit rules, then sends n8n a clean payload for AI summary, CRM and email.
The booking engine reads a mock JSON calendar and validates check-in, check-out, nights, capacity, blocked dates, minimum stay and accommodation type.
The demo combines accommodation, guests and extra services. Extras can be priced per booking, per guest, per night or per guest per night.
n8n validates the payload, generates a structured AI summary, creates or updates the contact, opens a deal, creates HubSpot tasks and notes, and sends a cautious email confirmation.
Architecture
Static Luna Blu Retreat interface with date search, accommodations, extras and guest data.
JavaScript validates calendar, availability, estimated price and the payload prepared for n8n.
The webhook validates input, enriches the request and produces an AI summary in structured JSON.
HubSpot receives contact, deal, task and notes; Resend sends confirmation that the request was received.
Proof of work
The project combines testable frontend logic and server-side automation. This reduces the risk of sending incomplete requests, inconsistent prices or context-free leads to the CRM.
Date, night, availability and pricing logic lives in a dedicated JS module, not scattered across the UI.
Accommodations, blocked dates, thresholds, extras and policies live in JSON for repeatable and testable demos.
The quote combines nightly rate, nights, guests and extras with different pricing modes.
The frontend sends guest, booking, extras, estimated total, privacy consent, source and timestamp.
OpenAI produces request type, lead priority, value range, suggested follow-up and CRM note.
The suite covers 9 cases: dates, nights, blocked dates, capacity, minimum stay, extra pricing and payload.
Beyond the demo
Check-out equal to or before check-in, past check-in and stays below minimum nights are blocked.
A night inside blockedDates makes the accommodation unavailable, while check-out only on a blocked date remains valid.
Requests above configured capacity are excluded before they are sent to the workflow.
Services with unavailableDates are not included in the quote for those dates or are routed to review.
The n8n webhook validates privacy_consent before proceeding with CRM and email.
Unparseable AI output, unavailable HubSpot or failed email lead to logs and manual review.
Stack
The stack is designed for a realistic but lightweight demo: static landing, testable booking engine, importable n8n workflow, HubSpot CRM and email provider.
HTML, CSS, vanilla JavaScript, booking form, hash routing, accommodations, extras and quote summary.
Mock calendar JSON, date validation, availability check, capacity, minimum nights, extra pricing.
n8n webhook, payload validation, OpenAI structured summary, lead priority and manual review.
HubSpot Contact, Deal, Task, Note, Resend email confirmation, final request log.
Need a similar system?
The demo uses a fictional retreat, but the same pattern works for hotels, B&Bs, spas, professional studios, local tour operators, rentals, consultants and services that require availability or a quote.
Prefer email? Reach me at luigi.scorzelli87@gmail.com