Blueprint Scanner
Public URL or HTML → structured frontend reverse-spec (Blueprint JSON).
A high-signal snapshot of a site's public UI — design tokens, structure, forms, tech stack and WordPress/Elementor/JetEngine architecture — for rebuilds, audits, migrations and design extraction.
It does not clone backends, databases or secrets.
Repo: youh4ck3dme/EEEEEE
What it does
- Scans a public
https:// page (or pasted HTML) into a versioned Blueprint JSON
- Same-origin multi-page crawl (1–20 pages) with partial recovery and transient retry
- Fetch chain: headless Chromium (Playwright) → static HTTP → archive.org Wayback fallback
- Extracts design tokens, CSS variables, typography, forms, headings, links, tech signals, assets, OG / JSON-LD metadata
- Optional WordPress / JetEngine / Elementor extraction: Elementor sections and globals, JetEngine dynamic fields and listing grids, REST/CCT hints, sitemap URLs
- Compiles an importable Elementor template JSON (v0.4 schema) from a blueprint
- AI Rebuild Studio: generates a system/user prompt +
tailwind.config.js fragment and an SPA-aware Architecture Compiler prompt (copy-only — the app makes no LLM API calls and needs no AI key)
- History vault, side-by-side compare, JSON / ZIP / Elementor export
- PWA & Asset Studio: ingest a public GitHub repo (or a master image) → icon set,
favicon.ico, manifest.webmanifest, HTML/Next.js head snippets, live preview and ZIP bundle — generated fully in the browser (Canvas + JSZip)
What it does not do
- Clone server code, databases,
.env, auth secrets or private APIs
- Bypass login walls or capture session cookies
- Scan private networks — localhost, RFC1918 and cloud-metadata addresses are blocked (SSRF guard)
- Promise 1:1 full-stack replication (target: ~70–90 % of the public frontend)
Features
| Area | Capabilities |
|---|
| Ingest | URL scan · HTML paste · baseUrl for relative paths |
| Fetch chain | Headless (Playwright) → HTTP static → Wayback |
| Crawl | Same-origin BFS · maxPages 1–20 · per-URL fault isolation |
| Resilience | Partial scanStatus · AbortSignal cancel · HTTP retry with backoff · thin-HTML / SPA shell detection |
| Extract | Meta/SEO · headings · links · forms · tech detection · design tokens · CSS vars |
| WordPress | JetEngine dynamic-field catalog · listing grids · Elementor sections/globals · template JSON |
| Hardening | SSRF blocklist · asset caps (10 MB/file, 50 MB total, 40 files, 12 s timeout) · browser process shield · withApiGuard |
| AI | Rebuild prompt + Tailwind config fragment · Architecture Compiler prompt (copy-only) |
| PWA Studio | GitHub repo analysis · icon engine · favicon.ico · manifest · head snippets · ZIP |
| Storage | Browser vault (last 15 scans) · server DB (Neon Postgres or embedded PGlite) · JSON import |
| Auth | Better Auth (federated OAuth via broker: Google, X) · can be disabled |
Tech stack
TanStack Start (React 19, file routes, server functions) · Vite + Nitro · Tailwind CSS 4 · Radix/shadcn UI · Zod · Playwright · node-html-parser · Better Auth + Kysely · Neon/pg or PGlite (Postgres in WASM) · JSZip · Vitest + Testing Library.
Quick start
Requires Node 22.
git clone https://github.com/youh4ck3dme/EEEEEE.git
cd EEEEEE
npm ci
npx playwright install chromium # needed for headless render; otherwise HTTP/Wayback fallback is used
npm run dev
Open http://localhost:8080, paste a public URL → Vytvoriť blueprint (the UI is in Slovak).
No .env is required for local development: without DATABASE_URL the app runs on an embedded PGlite database and applies migrations/ on startup.
Configuration
| Variable | Required | Purpose |
|---|
DATABASE_URL | prod | Neon / Postgres connection. Unset → embedded PGlite. Migrations run on npm run build |
VITE_AUTH_ENABLED | no | false disables sign-in; server functions use a dev user |
BETTER_AUTH_URL, BETTER_AUTH_SECRET | prod (auth) | Better Auth base URL and secret |
GROK_AUTH_ISSUER, GROK_AUTH_CLIENT_ID, GROK_AUTH_CLIENT_SECRET | prod (auth) | OAuth broker issuer and per-app client credentials |
PLAYWRIGHT_CHROMIUM_EXECUTABLE_PATH | no | Custom Chromium binary (e.g. serverless) |
PREVIEW_THUMBNAIL_TIMEOUT_MS | no | Headless thumbnail timeout |
SMOKE_BASE_URL, SMOKE_OUT_DIR, SMOKE_TIMEOUT_MS, BROWSER_SMOKE_TIMEOUT_MS | no | Smoke test runner |
Scripts
| Script | Description |
|---|
npm run dev | Vite dev server on 0.0.0.0:8080 |
npm run build | Production build, then db:migrate |
npm run build:dev | Vite build (development mode) |
npm run preview | Preview production build on 0.0.0.0:8080 |
npm run typecheck | tsc --noEmit |
npm run lint | ESLint |
npm run format | Prettier write |
npm test | Vitest run (default suite) |
npm run test:unit | Unit + integration (tests/unit, tests/integration) |
npm run test:smoke | Playwright UI smoke (tests/smoke/smoke-runner.mjs) |
npm run test:all | test:unit then test:smoke |
npm run db:migrate | Apply migrations/*.sql to DATABASE_URL (skipped when unset) |
Server API
TanStack server functions in src/lib/blueprint/server.ts (details in API):
| Function | Method | Purpose |
|---|
scanBlueprint | POST | Scan URL/HTML → blueprint (options: maxPages, render, wayback, captureAssets, wpJetEngine) |
getBlueprint | GET | Load a saved blueprint |
listBlueprints | GET | List saved blueprints |
deleteBlueprint | POST | Delete a blueprint |
compareBlueprintPair | POST | Diff two blueprints |
Exports
| Format | How |
|---|
| Blueprint JSON | Copy / download from the result view · exportBlueprintJson |
| ZIP archive | Captured assets + JSON · exportBlueprintZip |
| Elementor template | elementor-template-import.json (v0.4 schema) |
| AI Rebuild prompt | Clipboard · system + user prompt + Tailwind config fragment |
| Architecture prompt | SPA-aware UI rebuild brief (copy-only) |
| PWA bundle | ZIP with icons, favicon.ico, manifest and head snippets |
Project structure
src/
routes/ # TanStack Start routes (single-page app: index.tsx)
components/blueprint/ # scan-form, blueprint-view, history-list, compare-panel
components/pwa/ # PWA & Asset Studio
components/ui/ # shadcn-style primitives
lib/blueprint/ # scan pipeline, crawl, render, Wayback, design system,
# tech detection, WordPress/JetEngine, Elementor compiler, DB store
lib/scanner/ # browser shield, asset guards, pipeline, API errors
lib/ai-rebuild/ # rebuild prompter + architecture compiler
lib/pwa/ # GitHub repo analysis, icon/manifest engine
lib/auth/ # Better Auth server/client, middleware, PGlite dialect
lib/seo/ # JSON-LD helpers
lib/db.ts # Neon ⇄ PGlite switch
migrations/ # 0001_auth.sql, 0002_blueprints.sql
tests/ # unit/ · integration/ · smoke/ · fixtures/ · helpers/
scripts/ # migrate, browser-smoke, preview-thumbnail, qa-scan
docs/ # Architecture, Development, Testing, API, Security, Deployment, Runbook
.github/workflows/ci.yml # typecheck + unit + build, then Playwright smoke
Testing & CI
- 31 test files / 234 tests (unit + integration) — passing,
tsc --noEmit clean (verified 2026-10-24)
- CI (
.github/workflows/ci.yml, Node 22): job 1 typecheck → test:unit → build; job 2 installs Chromium and runs the blocking browser smoke test
Limitations
- Thin HTML / SPA: without a successful headless render, the shell may be empty (flagged via
isThinHtml)
- Login walls: only the public DOM is visible
- Asset budgets: max 10 MB per file, 50 MB total, 40 assets per scan
- Crawl size: max 20 same-origin pages per scan
- Serverless headless: Playwright needs a Chromium binary — see Deployment
- PWA Studio GitHub ingest uses the unauthenticated GitHub API (public repos only, 60 requests/hour per IP)
src/lib/multiplayer/ (WebRTC P2P client) is a scaffold — not wired into the UI and has no signaling endpoint yet.grok/skills/ (auth, neon, design-ui, multiplayer-p2p) and AGENTS.md are Grok Build agent instructions from the scaffolding template; they are not part of the runtime app
Documentation
License
Private — all rights reserved.
