Digitup Portlet-based builder
Internal interface to compose landings by combining reusable, configurable blocks.
Tourism
Dynamic landings
AI
RAG
Auto translation
Portlets
Mundomar Cruceros: landing builder and AI for tourism
Internal system to build custom commercial landings, translate pages with AI and improve cruise search via RAG and conversational assistants.
Project overview
Mundomar Cruceros is a cruise travel agency working with multiple cruise lines, international markets and ongoing commercial campaigns. Digitup built an internal platform focused on creating custom commercial landings: a builder based on reusable portlets that lets marketing and product teams publish pages without traditional page-by-page development.
The system goes beyond a classic CMS. It combines dynamic HTML, CSS and JavaScript generation, Gulp-based compilation, Apache Velocity (.vm) template rendering and an AI layer to translate full HTML pages, search ships and cruises via RAG and assist visitors with a recommendation chat. The result is a flexible foundation for digital tourism that speeds up campaigns and removes technical bottlenecks.
Context and need
In the cruise sector, every line, market and country requires different landings: specific offers, adapted commercial tone, images, pricing, routes and calls to action. Mundomar handles a high volume of campaigns and must respond quickly when a promotion, new ship or seasonal opportunity appears in a specific market.
Until then, creating or adapting landings meant relying on the technical team for every content, style or language change. That slowed commercial operations and made scaling to new countries harder. Cruise search across large catalogs did not always return useful results for end users, and translating full pages while keeping HTML intact was a costly manual process.
The problem to solve
The main challenge was twofold: give Mundomar an internal system capable of building commercial landings in a modular, reusable and configurable way without rewriting code for every campaign; and embed artificial intelligence where it adds the most value — page translation, semantic cruise search and conversational visitor assistance.
The solution had to address technical dependency on repetitive tasks, the complexity of adapting content to different lines and markets, and friction between commercial and development teams. It needed to be stable in production, keep visual consistency across landings and allow portlets, styles and AI features to evolve without rebuilding the platform from scratch.
The solution built by Digitup
Digitup designed and implemented a landing builder based on portlets: reusable content blocks configured per cruise line, campaign, market or language. Each portlet encapsulates logic, presentation and editable parameters so internal teams can compose full pages by combining modules without touching the system core.
Asset generation is dynamic: HTML, CSS and JavaScript are produced from configurations and Apache Velocity .vm templates, with Gulp as the compilation pipeline. On top of that base, three AI capabilities were integrated: automatic translation of HTML pages preserving structure and markup, cruise and ship search via RAG over the catalog, and an assisted chat that guides users toward relevant options based on their preferences.
What we built
An internal modular landing ecosystem with AI integrated into translation, search and assistance.
Dynamic generation
HTML, CSS and JavaScript generated from configurations, not isolated static pages.
Velocity engine (.vm)
Apache Velocity templates to render final pages with campaign logic and variables.
AI HTML translation
Automatic translation of full pages while keeping structure, classes and markup intact.
Cruise RAG
Semantic search over ship and route catalogs for more relevant results.
Recommendation chat
Conversational assistant that guides visitors toward cruises matching their criteria.
Key features
- Commercial landing builder
- Reusable, configurable portlets
- Per-campaign content configuration
- Style configuration per line or market
- Dynamic HTML, CSS and JS generation
- Apache Velocity (.vm) rendering
- Gulp compilation and bundling
- AI automatic HTML page translation
- RAG for cruise and ship search
- Assistance and recommendation chat
- Adaptation to lines, countries and languages
- Fast seasonal campaign publishing
- Extensible modular architecture
Artificial intelligence layer
AI was not added as a disconnected extra: it was integrated at three specific points in the workflow. HTML translation lets teams localize entire landings into new languages without redoing layout or breaking CSS classes, accelerating international expansion. Generative models work on existing markup, not loose text, preserving visual consistency across versions.
RAG over the cruise catalog improves search when users ask in natural language — destinations, duration, cruise line, ship type — instead of relying only on rigid filters. The assisted chat complements that layer: it converses with visitors, gathers preferences and recommends options from the catalog. Three distinct tools, one dynamic content architecture behind them.
Architecture and technology
The system is built on well-structured classic web: HTML5, CSS3 and JavaScript as presentation and interaction layers, with Gulp orchestrating asset compilation and Apache Velocity (.vm) rendering final pages from portlet data and configurations. The modular architecture allows new portlet types, style variations or campaign rules without touching the core.
LLMs and RAG pipelines connect on top of that stack for translation, semantic catalog indexing and chat responses. Separating content generation, rendering and the AI layer makes each part evolve independently: new markets, new lines or model improvements without rebuilding the full platform.
HTML5
CSS3
JavaScript
Gulp
Apache Velocity
LLMs
RAG
Generative AI
Modular architecture
Where Digitup contributed
System design, portlet development, compilation pipeline and applied AI integration.
Modular architecture
Design of the reusable, extensible portlet-based builder.
Portlet development
Implementation of configurable blocks for content, styles and campaigns.
Gulp + Velocity pipeline
Asset compilation and .vm template rendering in production.
AI translation
Integration of generative models to localize HTML without losing structure.
RAG and search
Semantic catalog indexing for natural-language queries.
Assisted chat
Conversational assistant connected to the catalog to recommend cruises.
Project impact
The platform let Mundomar create and publish commercial landings with much greater agility. Internal teams can compose pages for different lines, markets and campaigns without waiting for full development cycles, reducing technical dependency on repetitive tasks and speeding commercial response to seasonal opportunities or time-sensitive promotions.
The AI layer delivered value on three measurable fronts: faster translation into new languages, more useful cruise search for visitors and conversational assistance that turns vague queries into concrete recommendations. Mundomar now has an internal foundation ready to scale digital content in tourism — not isolated landings that must be rebuilt one by one.
Related services
Related projects
Need a landing or dynamic content system with AI?
We can help you design and build an internal platform with portlets, dynamic generation, automatic translation, RAG and conversational assistants tailored to your industry.