SearchLens — SEO / GEO / AEO Audit

A single-page tool that scans a web page for classic SEO, technical, and AI-search (GEO/AEO) readiness — with a 0–100 score, per-check evidence, ready-to-paste fixes, and PDF/CSV export.

• Frontend (index.html) — the full site + audit engine, runs client-side.
• Backend (api/scan.js) — a tiny serverless function that fetches any public URL server-side (bypassing browser CORS) so the "Enter URL" scan works live.

Two input modes:

• Paste HTML — works entirely in the browser (no backend needed).
• Enter URL — calls /api/scan, which fetches the page + its robots.txt + llms.txt and scans them.

---

Before you deploy — one step

Find-and-replace YOURDOMAIN.com with your real domain across the project (it appears in index.html, robots.txt, sitemap.xml, and llms.txt). That's the only edit needed. If you're only using the free *.vercel.app URL for now, use that instead.

Deploy to Vercel

Pick any one method.

A) Drag & drop (easiest)

1. Go to https://vercel.com/new.
2. Drag this whole searchlens folder onto the page (or zip it and upload).
3. Click Deploy. Done — you'll get a live URL.

B) Vercel CLI

npm i -g vercel        # once
cd searchlens
vercel                 # follow prompts (accept defaults)
vercel --prod          # promote to production

C) GitHub → Vercel

1. Push this folder to a GitHub repo.
2. In Vercel: Add New… → Project → Import the repo.
3. Framework preset: Other (no build step needed). Deploy.

No environment variables, no build command, no dependencies — it just works.

---

Structure

searchlens/
├── index.html        # frontend + audit engine (client-side) + full SEO meta & JSON-LD
├── api/
│   └── scan.js       # serverless URL-fetch proxy (bypasses CORS)
├── robots.txt        # crawlers + AI crawlers allowed, links sitemap
├── sitemap.xml       # add a <url> per new page you build
├── llms.txt          # AI-readable map of the site
├── favicon.svg       # scalable icon (+ favicon.ico fallback)
├── favicon.ico
├── og-image.png      # social/AI share preview (1200×630)
├── package.json      # Node >=18 (uses built-in fetch, no deps)
├── vercel.json       # function config
└── README.md

How it works

• index.html runs all deterministic checks in the browser (title, meta, headings, schema, AI-crawler rules, llms.txt, answer-first, etc.).
• For Enter URL, the browser can't read another site's HTML (CORS), so it calls GET /api/scan?url=…. The function fetches the page + robots.txt + llms.txt on the server and returns them; the browser then scans that HTML with the same engine.
• api/scan.js has basic SSRF protection (blocks localhost / private IPs), a 10s timeout, and a size cap.

Cost

Runs on Vercel's free (Hobby) tier: static hosting is free, and the function fits comfortably in the free serverless allowance. Effectively $0/month at low–moderate traffic.

Next layers (optional)

• AI-written fixes — send the page to a cheap/free LLM (Gemini Flash, Groq, or the user's own key via BYOK) and return page-specific rewrites. Add a second function (api/fix.js) and keep the API key in a Vercel Environment Variable — never in the frontend.
• Backlinks / authority — via DataForSEO (reseller-friendly) using BYOK or a paid tier; keep it out of the free path.

Notes

• Opening index.html directly from disk (or in a preview host) works for Paste HTML; Enter URL only works once deployed (it needs /api/scan).
• The score is a diagnostic heuristic, not a guarantee.