Skip to content
Get Started. Free Consult
Free tool · results on screen in seconds · no code access

Find what your AI coder left exposed.

Tools like Lovable, Bolt, Cursor and Claude Code ship fast, but nearly half of AI-generated code carries a security flaw. Paste a URL and Vibe Scan checks for leaked API keys, open databases, missing headers and the mistakes vibe-coded apps make most.

vibe-scan·~/scans/your-app
Target URL
Free · no code access · publicly visible signals only · results on screen instantly
01Coverage

Six layers, one public URL.

A full front-end security pass over the vulnerabilities AI-generated apps introduce most. No source code, repo or database access required, just a URL we can reach.

01Headers

Security headers

The HTTP response policy layer.

Content-Security-PolicyHSTSX-Frame-OptionsReferrer-PolicyPermissions-PolicyCORS config
02Secrets

Exposed keys

Credentials leaked into shipped JS.

OpenAISupabase service-roleStripeAWSFirebaseGeneric tokens
03Client

Client-side security

What runs in the visitor's browser.

Outdated libs / CVEsMissing SRIMixed contenteval() usageXSS vectors
04Vibe

Vibe-code patterns

Mistakes specific to AI builders.

Supabase misconfigOpen tables / no RLSExposed source maps__NEXT_DATA__ leaksDebug routes
05Paths

Exposed files

What's reachable that shouldn't be.

.env.git/admin/api/docsBackup files
06Infra

Infrastructure

Transport and config hygiene.

HTTPS enforcementServer fingerprintCookie flagsStack detection
02Process

Paste, scan, act.

01

Paste your URL

Any deployed app with a public address, Lovable, Bolt, Vercel, Netlify, anywhere. No install, no code access.

02

We scan live

Headers, JavaScript bundles, Supabase config, exposed paths and known CVE patterns, all checked in seconds against your real URL.

03

Get a fix list

Every finding with a severity, plain-English impact and the exact change to make, graded and shown on screen the moment the scan finishes.

03Why

Speed without a security review.

The trade-off

The review your AI tool skipped.

Lovable, Bolt, Cursor, Replit and Claude Code make building astonishingly fast. But they routinely introduce flaws a developer would catch in code review, and most vibe-coded projects never get one.

Vibe Scan is that missing review: the security pass your AI coder should have done, but didn't.

The usual suspects

What we find almost every time.

  • Service-role keys shipped in client JSvery common
  • Missing Content-Security-Policyvery common
  • Hardcoded OpenAI / Stripe / AWS keyscommon
  • .env reachable at the site rootcommon
  • CORS allowing any originfrequent
  • Public tables with no RLS policyfrequent
04Triage

What to do with the results.

Critical · right now

Rotate every exposed credential.

Service-role keys, database passwords, OpenAI and Stripe secrets, rotate them immediately. Every minute they stay public is an open door.

High / Medium · this week

Close the attack surface.

Missing headers, outdated libraries, exposed admin routes. None are a breach alone, but each is a step an attacker can chain into one.

Want a human to look?

Get a manual audit.

We go past the automated scan, reviewing source, database rules, auth logic and API endpoints across Next.js, React, Svelte and Nuxt.

Request audit
05FAQ

Vibe Scan, answered.

An automated check for apps built with AI coding tools like Lovable, Bolt, Cursor, Claude Code and Replit. It looks for the vulnerabilities those tools introduce most: exposed API keys, missing security headers, open database access, leaked credentials in JavaScript bundles, and misconfigured Supabase or Firebase.

No. It only reads publicly visible signals: HTTP response headers, HTML source, external JavaScript bundles, and whether common sensitive paths are reachable. It never touches your repository, database, or any private resource.

Any app with a public URL. It is tuned for apps built with Lovable, Bolt.new, Cursor, Claude Code, Replit, Windsurf, v0 and Copilot, plus framework-specific issues in Next.js, React, Vite, Nuxt and Svelte.

Yes, no credit card and no trial period. Results appear on screen the moment the scan finishes. We ask for your name and email so VibeZero can follow up if you want help fixing what is found. We will not spam you.

It reliably catches the most common and dangerous issues: exposed secrets, missing security headers, vulnerable libraries, open database tables, misconfigured paths. It cannot see server-side code, database security rules or authentication logic. For that, request a manual audit.

The scanner only reads public information, so technically yes. Please only scan apps you own or are authorised to test.

Most scans finish in 5 to 15 seconds. Larger apps with many JavaScript files can take up to 20.

Rotate exposed credentials immediately, then prioritise high and medium findings within the week. If you would like help, request a free manual audit and we will go deeper.

Ship fast. Ship secure.

Scan my vibe-coded appRequest a free code audit

instant · results on screen · no code access needed