Plug ScamShield into the places scams travel

Telcos can score SMS before delivery, banks can check a payee link before a transfer, delivery companies can verify links in customer complaints, and messaging apps can warn inside the chat.

Try it

{ }

Check a link (with live checks)

GET /api/v1/check?url=https://salik-uae-pay.top/settle
{ "verdict": "dangerous", "score": 93, "domain": "salik-uae-pay.top", "impersonates": "salik",
  "reasons": [{ "id": "brand_in_domain", "en": "Uses the name “Salik” but is NOT an official Salik domain…", "ar": "…" }, …],
  "ageDays": 3, "ms": 812 }

Check a message (rules only, < 5 ms, no network)

GET /api/v1/check?text=…&sender=+447700900123

Built for inline SMS-firewall use: deterministic, explainable, Arabic + English, no external calls.

Full scan (what the app uses)

POST /api/scan
{ "type": "message" | "url", "input": "…", "sender": "optional", "ai": true }

Adds redirect unrolling, RDAP domain age and registrar, landing-page inspection (card / OTP / password / Emirates ID forms, Telegram exfiltration, off-site form posts), community report counts, and the AI analyst’s second opinion.

Report (anonymous fingerprint)

POST /api/report
{ "fingerprint": { "brand": "salik", "playbook": "fake_fine", "domain": "salik-uae-pay.top", "tld": "top", "score": 93, "signalIds": ["brand_in_domain", "risky_tld"] },
  "channel": "sms", "emirate": "AD", "lostMoney": false }

Intelligence (police)

GET /api/stats?days=14&demo=0     aggregated campaigns, surges, top domains
GET /api/hunt?brand=rta            Certificate-Transparency + DNS look-alike hunt
GET /api/takedown?domain=x.top     evidence pack + drafted abuse notice
POST /api/hunt  (Bearer analyst key)  { "domain": "x.top", "status": "takedown_requested" }

Embed the engine

import { analyzeMessage, analyzeUrl } from 'https://scamshield-ai-eaj.pages.dev/engine/scamshield.js';
const r = analyzeMessage(smsText, { sender });   // { verdict, score, signals[{en, ar}], advice, … }

Rate limits on this prototype: 120 requests/min per client for /api/v1, 30/min for /api/scan.

← Back