Frontend Slides Generator

The Problem

When turning a talk, product review, or internal update into a browser presentation, the hard part is usually not producing an HTML file. It is avoiding dependency-heavy multi-file outputs, slides that scroll internally, noisy animation, template-like styling, or PPT conversions that leave notes, images, and slide order behind.

How It Works

fronted-slides treats an HTML deck like a small frontend project. The core constraints are explicit:

• Zero dependencies: it defaults to one self-contained HTML file with inline CSS and JS.
• Mandatory viewport fit: each slide uses 100vh / 100dvh, avoids internal scrolling, and splits content when it does not fit rather than shrinking text below readable sizes.
• Style previews first: it generates three single-slide previews so the user can choose typography, color, layout, and motion direction.
• Production-grade structure: semantic slide sections, CSS custom properties, keyboard, wheel, and touch navigation, IntersectionObserver reveal triggers, and prefers-reduced-motion support.

For PPT / PPTX files, it prefers python-pptx to extract text, images, and speaker notes, then follows the same style discovery and build workflow while preserving order and assets.

Boundaries And Caveats

It is well suited to talks, reviews, teaching, and product launches that need a visually coherent HTML deck. It is not intended for complex interactive applications, collaborative editing, or server-rendered systems. Content density is intentionally limited, so long paragraphs, dense bullets, and oversized code blocks should be split rather than crammed into a single slide.