Features · How it works · Why Pagelight · Local Development
Pagelight AI reads your article, extracts the key data, and renders it as a beautifully designed piece with animated charts woven throughout the text.
- Visualizations: Animated charts and graphs
- Editorial Design: Custom aesthetic per article
- Woven Narrative: Visuals placed between text
At its core, Pagelight uses the Claude Sonnet 4.6 LLM to analyze the structure and content of an article, blog post or report. It identifies key data points, trends, and narrative elements, then picks pre-built visualizations that complement the text.
Finally it combines the analyzed data and selected visualizations into a custom page that weaves visuals between paragraphs, creating a more engaging reading experience
Before using Pagelight you need to provide a Anthropic API key. Your API key will be stored locally in your browser and is never sent to our servers or anyone else except Anthropic.
This started as a fun experiment: using AI to spice up long-form reading by pulling out the “interesting bits” and giving them a visual home. Articles are full of numbers, comparisons, and timelines — Pagelight tries to make those moments feel tangible.
Initially built for personal use, I soon realized it could be useful for others too. It’s a way to make dense information more engaging and accessible, especially for people who are visual learners or just want a fresh way to experience content.
Pagelight AI is built with Next.js and Bun. To run it locally, make sure you have Bun installed, then clone the repo:
git clone https://github.com/betahuhn/pagelight.gitInstall dependencies with Bun:
bun installStart the Next development server:
bun run devEdit app/page.tsx to customize the homepage content. The app will automatically reload as you make changes.