PAGE2IMG

Beautiful pages
from any document

AI agent plugin that reads any file or content, generates a mobile-optimized webpage with your chosen design style, and captures a high-resolution long screenshot. Zero config. Works offline.

claude plugin marketplace add laosuan/page2img
claude plugin install page2img@page2img

5 design styles

McKinsey, Minimal, Dark Tech, Academic, Startup. Professional output every time.

3x retina capture

1170px wide HD screenshots. Tile-based stitching for unlimited page height.

Multi-platform

Native plugins for Claude Code, OpenCode, and OpenClaw.

FEATURES

Persistent beauty,
zero plumbing

One command turns raw content into polished output. No design skills, no build steps, no API keys required.

01

Mobile-first design

Every page is optimized for 390px viewport with touch-friendly elements, CJK font support, and responsive breakpoints.

02

Smart content parsing

Markdown, plain text, structured data — the agent identifies sections, headings, metrics, and creates semantic HTML automatically.

03

HD long screenshots

Full-page 3x retina capture using Chrome DevTools Protocol. Tile-based stitching bypasses GPU texture limits for any height.

04

Fully offline

No API keys, no server, no network required. Uses your local Chrome for screenshots. Self-contained HTML with inline CSS.

STYLES

Pick your aesthetic

Five carefully designed styles for different contexts. Each includes typography, color palette, and a full component library.

McKinsey

Consulting report

DEFAULT

Minimal

Clean whitespace

Dark Tech

Terminal mode

Academic

Paper aesthetic

Startup

SaaS landing

PLATFORMS

One plugin.
Every agent.

Native integration for the three major AI coding agent platforms. Same capability, platform-native interface.

Claude Code

Hooks + Skills plugin. Use /page2img in any conversation. Auto-hints via UserPromptSubmit hook.

$ claude plugin marketplace add laosuan/page2img
$ claude plugin install page2img@page2img
OpenCode

TypeScript SDK plugin. Exposes page2img_convert and page2img_styles tools with system prompt injection.

$ npm install github:laosuan/page2img#main
OpenClaw

Extension with openclaw.plugin.json manifest. Configurable default style, Chrome path, and scale factor via plugin config UI.

$ npm install github:laosuan/page2img#main
HOW IT WORKS

Four steps. Seconds.

From raw content to polished deliverable.

01

Read content

Your AI agent reads the source file, pasted text, or fetched URL. Markdown, plain text, structured data — anything works.

02

Generate HTML

Creates a single self-contained HTML file with inline CSS using your chosen design style. Mobile-first, semantic, beautiful.

03

Capture screenshot

Puppeteer launches Chrome headless at 3x retina (1170px wide). Full-page capture with tile-based stitching for unlimited height.

04

Deliver output

You get index.html (open in browser) and screenshot-hd.png (share anywhere). Clean, no leftover dependencies.

FAQ

Questions & answers

Does it need an API key or server?
No. Everything runs locally using your installed Chrome. No accounts, fully offline.
How large can the screenshots be?
Unlimited height. Tile-based capture works around Chrome's GPU texture limits. Width is 1170px (3x retina of 390px mobile viewport).
Does it support Chinese / CJK content?
Yes. All styles include CJK font fallbacks and proper line-breaking. The McKinsey style was originally designed for Chinese reports.
Can I create custom styles?
Ships with 5 built-in styles. Since output is plain HTML/CSS, you can edit the generated file or ask your agent to customize further.
What if Chrome isn't installed?
HTML generation works without Chrome. For screenshots, you need Chrome or Chromium. The plugin auto-detects common paths on macOS, Linux, and Windows.