FvRE v3.0 — Universal Scanner (results)

Tested the v3.0 scanner against 5 sites: 2 WordPress/Elementor, 1 Next.js, 1 static HTML, 1 minimal example. All classified correctly. Recipes in JSON.

Test matrix

URL	Title	Classification	Confidence	Assets	Interactions	JSON
wp-oscar.freshvibeapps.com	Oscars Tree Academy	elementor	100%	28	10	oscar-wp.json
hopfan.freshvibeapps.com	Hopfan Europe	elementor	100%	70	11	hopfan-wp.json
react.dev	React	nextjs	60%	40	0	react-dev.json
iana.org	IANA	unknown	0%	—	—	iana-static.json
example.com	Example Domain	unknown	0%	0	0	example-com.json

What the recipe contains

Every recipe has the same shape regardless of source:

• source — URL, timestamp, scan duration
• pages — the rendered page(s) with full DOM tree
• page_list — internal links found (crawlable surface)
• assets — images, fonts, videos (deduplicated by URL hash)
• interactions — forms, modals, carousels, tabs
• navigation — extracted nav structure
• design — colors, fonts, spacing, radii from CSS
• classification — platform guess + confidence + evidence

How to use

python3 fvre_v3_scan.py <url> --out recipe.json --max-pages 20

Scanner is ~600 lines of Python, single file, no install (just playwright + Python 3.11+). Reads a URL, renders with Playwright, outputs the generic recipe schema from the design doc.

Next: v3.1 (recipe → FVS app)

Now that we can scan anything, the converter takes a recipe.json and produces a runnable FVS app. Three modes:

• Static mirror — render once, serve as-is
• Live edit — generate editable FVS pages
• Hybrid — static shell + editable zones

← back to v3 design · ← back to index