url2svg Roadmap

Phase 1 — Core capture (done)

• URL → headless Chrome → JS extraction → VisualElement tree → SVG
• Backgrounds, borders, border-radius, opacity, overflow clipping
• Text runs with font properties, text decoration
• Images as embedded data URLs
• Box shadows
• Links (href)
• Full-page scroll capture mode

Phase 2 — Output modes & file size

• Keep current "full" mode (large, high-fidelity files)
• New compact default mode
  • Target 1024x800 viewport
  • Downscale / re-encode embedded images (e.g. JPEG quality reduction, resize to rendered dimensions)
  • Strip redundant precision from coordinates (round to 1 decimal)
  • Collapse empty/invisible elements
• CLI flag: -mode full|compact

Phase 3 — Fidelity improvements

• Gradients (linear-gradient, radial-gradient backgrounds)
• Pseudo-elements (::before, ::after)
• SVG/canvas element passthrough
• Web font embedding (subset and inline)
• CSS transforms (translate, rotate, scale)
• Form controls (inputs, buttons, selects as visual approximations)

Future ideas

• Streaming / incremental output for large pages
• Diff mode: capture two URLs and highlight visual differences
• PDF output via SVG → PDF conversion
• WASM build for in-browser use (no server-side Chrome needed)