AI-Powered Website Builder
Kilat.store
An AI-powered website builder allowing Indonesian small-medium businesses (UMKMs) to construct their business landing pages in 60 seconds.
Outcome
Successfully built a multi-tenant platform with visual editor capabilities and automated generation pipelines, lowering the digital commerce entry barrier.
Problem
Indonesian MSMEs lack the technical resources to design and build websites, while generic builders are too complex or expensive.
Approach
Designed a modular, JSON-based SSoT (Single Source of Truth) block schema shared between the Next.js visual editor and the backend AI generation pipeline.
Architecture
A monorepo comprising an Astro marketing site, Next.js editor UI powered by Puck Editor, Vite admin portal, and Ruby on Rails 8 REST API handling background jobs (Solid Queue/Solid Cache) for DashScope/BytePlus AI pipelines. Wildcard client domains map to static HTML via DCDN.
Result
Enabled seamless generation of high-quality landing pages with microsecond rendering speeds across Indonesia.
Lessons learned
Sharing block specifications between the AI prompt templates and editor schema simplifies synchronization when introducing new UI modules.
Constraints
- Must run smoothly on low-end mobile devices.
- Generation and editing flow must feel instantaneous under volatile cellular networks.
- Maintain extremely low hosting costs for client landing pages.
Technical decisions
- • Shared blocks.json registry as single source of truth.
- • Asynchronous AI generation via Rails Solid Queue to prevent timeouts.
- • Utilized Terraform to provision Alibaba Cloud ECI and DCDN for static site delivery.
Key features
- • 60-second AI landing page generation.
- • Interactive block-based Puck Editor canvas.
- • Automatic publishing to *.kilat.my.id subdomains.