Free full-page capture, markup, and PDF export for QA, design review, and frontend debugging

bicaos — Sun, 04 Oct 2026 17:09:22 +0000

If your work involves capturing, marking up, and sharing what a whole webpage actually looks like — for a bug report, a design review, or a QA record — PageStitch 0.3.0 adds the two things that usually sit behind a paywall in other capture tools: a free image editor and free PDF export.

Capture is unchanged: trigger it with Alt+Shift+S or the toolbar icon, and get one PNG of the full page, or of the one scrollable area (an internal list, table, or panel) that fills at least half of it when the page itself already fits the window. Pages too large for one PNG still export as lossless numbered parts rather than being shrunk or cropped — useful when a QA record or design handoff needs to stay pixel-accurate.

New in 0.3.0: click "Edit image" on the result page to crop, draw rectangles, ellipses, arrows, lines, freehand pen strokes, and highlights, add text or one of 40 emoji, and pixelate out sensitive details (an account number, a face, an internal URL) before you file the capture anywhere. "Add page info" stamps the site, title, and date onto the export for an auditable record; "Add watermark" adds custom text at a position and opacity you choose. Undo/redo, zoom, and one-key tool shortcuts are built in.

Export as PNG as before, or as a PDF: one long page, A4, or US Letter, with smart page breaks that look for a blank row near the page boundary so a cut doesn't land mid-table or mid-paragraph. A capture too large for a single PNG can still come out as one PDF.

On permissions and privacy: still exactly activeTab, scripting, and offscreen, with no account, no analytics, and no uploads, and no remote code — the editor and PDF export run locally, same as capture always has, which matters if your team has to justify extensions used against internal or client pages. Editor settings and your chosen PDF page size are remembered locally so you're not reconfiguring them every time. The UI is available in seven languages if your team or client base isn't all English-reading.

Known limits, unchanged: Chrome system pages, the Web Store, other extension pages, and non-HTTP(S) URLs can't be captured, the source tab needs to stay active during capture, pages that animate continuously or keep loading content as you scroll can still produce an imperfect stitch, and the exported PDF is image-only (no selectable or searchable text).

Get PageStitch free on the Chrome Web Store

I Built an AI That Roasts Websites Like Gordon Ramsay

bicaos — Fri, 13 Feb 2026 09:35:07 +0000

I got tired of generic website feedback. "Looks nice" doesn't help anyone.

So I built RoastMyWeb — paste any URL and AI brutally roasts your website with a score out of 100.

What it does

Crawls your site and reads the actual HTML
Analyzes design, content, SEO, performance, and UX
Delivers a savage but actionable roast
Gives you a prioritized list of fixes

The Stack

Next.js + Tailwind CSS (frontend)
Cheerio (HTML parsing)
Llama 3.3 70B via Groq (AI analysis — free tier!)
Vercel (hosting — free)

Total cost to run: $0/month

Try it

https://roastmyweb-tan.vercel.app

Fair warning: it doesn't hold back.

What would you want it to analyze deeper? Drop your score in the comments!

DEV Community: bicaos

Free full-page capture, markup, and PDF export for QA, design review, and frontend debugging

I Built an AI That Roasts Websites Like Gordon Ramsay