Full-Page Screenshot

Free

Captures full-page screenshots of web pages via Chrome DevTools Protocol with zero dependencies. *By [@LewisLiu007](https://github.com/LewisLiu007)*

FreeFree tier

Inputs: url

Type

Open Source

Visit Website GitHub (40 stars)

About Full-Page Screenshot

Full Page Screenshot is a zero-dependency Claude Code skill for capturing full-page screenshots of entire web pages as a single PNG image. It connects directly to Chrome DevTools Protocol (CDP) — no Puppeteer, Playwright, or npm install needed. The skill supports SPA detection, DOM stability waiting, lazy-loaded image triggering, and ultra-long page tiling (pages over 16,000px are captured in tiles and stitched with Python PIL). It can screenshot an existing browser tab by targetId or open a URL, capture, and close automatically. Designed to work within Claude Code, it can also be used as a standalone CLI tool. Requires Node.js 22+ and Chrome with remote debugging enabled. Licensed under MIT.

Key Features

Full-page capture renders entire scrollable page into one PNG

SPA support auto-detects overflow-y: scroll/auto containers and expands them

DOM stability wait monitors element count changes until SPA rendering completes

Lazy-load handling scrolls through page to trigger IntersectionObserver images

Ultra-long page tiling captures pages over 16,000px in 8,000px tiles and stitches with Python PIL

CDP Proxy fallback automatically falls back to proxy API when CDP proxy holds WebSocket

Two capture modes: screenshot existing browser tab by targetId, or provide URL to open, capture, then close

Zero external npm dependencies — uses only Node.js built-in modules (fs, net, path, os, child_process)

Configurable viewport width and device pixel ratio

Works as Claude Code skill or standalone CLI tool

Pros & Cons

Pros

Zero dependencies — no npm install, no Puppeteer, no Playwright required
Works seamlessly within Claude Code via natural language commands
Handles modern web complexities like SPAs, lazy-loading, and extremely long pages
Configurable viewport width and DPR for high-quality captures
MIT license allows free use and modification
Can be used standalone via CLI, not just within Claude Code

Cons

Requires Chrome browser with remote debugging enabled and manually configured
Requires Node.js version 22 or higher
Optional Python PIL dependency needed for stitching ultra-long pages (over 16,000px)
Ultra-long pages at high DPR may cause Chrome memory issues — recommends lowering DPR
Not a standalone application; tied to Claude Code ecosystem or CLI usage

Best For

Capturing full-page screenshots of websites for documentation or archivingTesting visual rendering of long pages, SPAs, or pages with lazy-loaded contentAutomating screenshot capture within Claude Code workflowsGrabbing high-resolution screenshots of wide tables or dashboards (viewport width adjustable)

FAQ

How do I enable Chrome remote debugging?

Open chrome://inspect/#remote-debugging and check 'Allow remote debugging for this browser instance'.

How do I install this skill for Claude Code?

Run 'npx skills add LewisLiu007/full-page-screenshot' or manually clone the repo into ~/.claude/skills/full-page-screenshot.

Can I use this without Claude Code?

Yes, it can be used as a standalone CLI tool. Use node scripts/full-page-screenshot.mjs with --url or targetId parameters.

Does it support single-page applications (SPAs)?

Yes, it auto-detects overflow-y: scroll/auto containers, expands them, and waits for DOM stability before capturing.

How are pages longer than 16,000px handled?

They are captured in 8,000px tiles and stitched using Python PIL. For very long pages, lowering DPR to 1 is recommended to avoid memory issues.