Security scanner for vibe-coded apps

Find what your AI-built app is leaking before someone else does.

SafeWeave checks your code and your live app for exposed keys, open databases and the bugs AI builders tend to ship. Every finding comes with a fix prompt you paste straight back into Lovable, Cursor or Claude Code.

Read-only · about a minute · free account

Free shows your grade and finding counts. Every finding with its fix prompt, and the live app checks, are on Cloud (14-day trial).

Or scan a public GitHub repo →
Or scan locally from your editor, no signup for the first 10 scans:
your-app.lovable.appExample report
D
Two critical issues are holding this app at a D.C at 70 · B at 80
  1. 01
    criticalSupabase service-role key in the JavaScript bundle
  2. 02
    criticalRow Level Security is off on profiles
  3. 03
    highNo Content-Security-Policy header
Fix prompt for 01

Create the browser Supabase client with the anon key, move code that needs the service-role key into a server function, and rename the env var so it no longer starts with VITE_.

Built for apps made withLovableBoltReplitv0Base44CursorClaude CodeWindsurf
What we check

Three places your app can leak. One report.

AI builders write a lot of code fast. The mistakes repeat: keys in the bundle, tables without RLS, auth checks that never run.

018

Your code

8 scanners on every push and pull request.

  • SAST · secrets · dependencies
  • IaC · container · license
  • DAST · posture
0237

Your live app

Read-only checks against the URL you deployed to.

  • Keys in the JavaScript bundle
  • Security headers and TLS
  • Exposed files and open data
0359

Your vibe stack

Rules written for the stacks AI builders reach for.

  • Supabase RLS and storage
  • Next.js, Firebase, Stripe
  • AI agents and MCP servers
How it works

Scan. Read one sentence. Paste the fix.

  1. 01

    Point it at your app

    Paste the live URL, connect the GitHub repo your builder syncs to, or run it in your editor with one command.

  2. 02

    Read the verdict

    A grade from A to F, one plain sentence on what is holding it down, and a fix queue with the worst first.

  3. 03

    Paste the fix, re-scan

    Copy the fix prompt into Lovable, Cursor or Claude Code. The next scan closes the finding by itself.

Pricing

Start free. Pay when you ship.

Free

$0

A first look at one app. Full findings, fix prompts and live app checks are on Cloud.

  • 10 local scans, no signup
  • 7 scanners locally with a free account
  • 1 repository with its grade and finding counts
Create a free account

Cloud

Most picked
$29/month

For one builder shipping for real.

  • 5 repositories
  • 15 cloud scans per repo a month
  • GitHub checks on pull requests
  • 59 vibe-stack rules + 37 live checks
Start 14-day trial

Cloud Plus

$59/month

For a small team with a few apps.

  • Everything in Cloud
  • 15 repositories
  • Unlimited cloud scans
  • Up to 5 people
Upgrade to Cloud Plus

Custom

Custom

For agencies, studios and larger teams.

  • Everything in Cloud Plus
  • Repositories sized to your team
  • Teammates sized to your team
  • A plan built around how you ship
Talk to us

Cloud plans include a 14-day free trial. Compare every feature

Questions

The short answers.

Does the live app scan change anything?

No. Live checks are read-only: SafeWeave loads your site the way a visitor would and never writes to your app or database.

Do you keep my code?

On Free, scanners run on your machine and your code stays local. On Cloud, files are sent over HTTPS for the scan and the checkout is deleted afterwards. Findings keep a short excerpt so you can fix them in context, and no excerpt is ever kept for a secret.

I didn’t write the code. Can I still fix things?

Yes. Each finding has a plain-English explanation and a fix prompt written for your AI builder, so you can ask it to make the change.

Which compliance frameworks are covered?

Findings map to OWASP Top 10, SOC 2, PCI-DSS, HIPAA, GDPR, ISO 27001 and NIST 800-53.

More questions

Know your grade before your users do.