Sample alert How it works Coverage PENTEST METHODOLOGY DOCS PRICING FAQ MCP CONTACT LOG IN SCAN MY APP →
LANGUAGE
Continuous security monitoring for Lovable, Bolt, Cursor, v0 and Supabase apps that change every day

AI keeps changing your app.
We keep checking it.

A one-time audit is out of date after the next prompt. We re-test your live app for open Supabase tables, leaked keys and exposed pages, compare it with the last scan, and alert you only to what’s new. A security engineer reproduces every finding.

14-day free trial · Read-only by default · Just your URL, no SDK · Turn monitoring off anytime

Free leak check, no signup ↗ · From the team behind the monthly Vibe Coding Security Report →

Why AI-built apps get scanned

UpGuard · Oct 2026
16,326
Supabase databases found with tables anyone could read
GitGuardian · Oct 2026
2×
Secret-leak rate in Claude Code-assisted commits vs. baseline (3.2% vs 1.5%)
Shippable · Oct 2026
29%
Of public Lovable repos had security issues in static checks
VibeEval study · 2026
Drift
Apps that were clean at launch regress as features ship: a new table without RLS, a new route without an ownership check. Read the study →

Third-party figures are vendor-reported and not reproduced by us. Sources in our October report →

Only what’s new

Each scan is compared with the last one. You hear about new exposures and fixes that broke again, not the same list every week.

Reproduced by a human

A security engineer reproduces each finding before delivery. You get evidence, not a wall of maybes.

Re-checked until fixed

Fix guidance ready for Claude Code or Cursor. The next scan confirms the patch held, and alerts you if it re-opens.

A regression alert,
start to fix.

A clear explanation of what is exposed, who it affects, and how to fix it.

NEW SINCE LAST SCAN · HIGHILLUSTRATIVE EXAMPLE

Private invoices are visible without signing in

Appeared after the “add billing page” change. Someone with an invoice link can see customer details, even when they are signed out.

Affected area
Customer billing
Issue
Missing access checks
EXPOSED CUSTOMER DATA
Customer invoicePrivate
Sign-in check missing

Private billing data. An unprotected page.

The issue

A signed-out visitor can open the invoice and read customer details.

Your next step

Require sign-in and check that the invoice belongs to the customer requesting it.

Retest the fix

Scan again to confirm that other visitors can no longer access the invoice.

Scan. Verify. Fix.

Initial triage starts the process. Deeper probing and human verification follow.

Add your app

Provide your URL and testing scope. No SDK or source-code changes needed.

Get initial triage

First triage in under 60 seconds. A deeper automated scan typically takes 3–8 minutes, depending on scope.

Check the evidence

An engineer reproduces findings and assesses impact before delivery. Verification time depends on the finding.

Make the change

Receive evidence and fix guidance. Use the next scan to check your patch.

Our methodology and limits →

Keep private things private.

Choose a security check to explore
ILLUSTRATIVE SCENARIO
Customer invoicePrivate
Account ownerAccess allowed
Other visitorAccess blocked

Cross-account access checks

We check whether someone can cross account boundaries and see records that should be private.

ILLUSTRATIVE SCENARIO
Your published app
Private credential exposed•••• •••• •••• ••••
Needs attention
Remove it from the app. Rotate the credential.

Private credentials in your frontend

Identify private credentials shipped with your frontend. Public keys alone are not a finding.

ILLUSTRATIVE SCENARIO
Your live app
Sign-in boundariesChecked
Private pagesChecked
Account permissionsChecked
A finding includes evidence and a next step.

Tests against the deployed app

We probe the deployed app, so findings reflect what a visitor can access.

4Browsers coveredChrome · Firefox · Safari · Edge

Use fix guidance in Claude Code or Cursor, with MCP and webhook integration for your team. Explore integrations →

Plans by apps, scans, and seats

Every plan includes live app testing and engineer-verified findings. Start with a 14-day free trial. A card is required; we email you 3 days before the first charge, and you can cancel in one click.

No verified finding during your trial? We extend it by 14 days, free. See a sample regression report ↑

ESTIMATE YOUR PLAN

Tell us what you ship. We pick the lowest plan that fits.

3
Scan cadence
1
ESTIMATED SCANS
12
scans / month · 3 apps, weekly
LOWEST PLAN THAT FITS
Pro
$25/mo

Daily security checks for your own apps.

12% of plan used100 ceiling
Start with Pro

Or pay $249 once for Lifetime.

SOLO BUILDERS
PRO
$25/mo
Daily security checks for your own apps.
  • Unlimited projects
  • Engineer-verified findings
  • Daily re-scans
  • 24h email support
Start free trial →
ONE-TIME PURCHASE
LIFETIME
$249once
Pro coverage with one payment and no renewal fees.
  • Everything in Pro
  • Weekly monitoring (early access)
  • 14-day money-back guarantee
  • Priority support
  • No renewal fees
Choose Lifetime →

14-day money-back guarantee · Cancel anytime on monthly plans · Questions: contact us →

Before you start.

Billing, access, and what we actually test.

Will I be charged when the trial ends?
The free trial lasts 14 days and includes the Pro feature set. A card is required to start. We email you 3 days before the first charge, and you can cancel in one click. If we don't find a verified issue during your trial, we extend it by 14 days, free.
How is this different from running ZAP or nuclei myself?
Generic scanners list what might be wrong. We test your deployed app for cross-account access, open Supabase tables and leaked keys, compare each scan with the last one, and a security engineer reproduces every finding before it reaches you.
Will you touch my code, database, or live data?
No SDK, source code, or database access is needed. You provide your app's URL and testing scope, and we test the deployed app the way a visitor would. Probes are read-only by default; destructive testing requires explicit opt-in.
Who verifies the findings?
The agent captures evidence, then a security engineer reproduces the finding and checks its impact before delivery. Initial triage is not the same as a human-verified report.
Will I get alert spam?
No. Each scan is compared with the last one, so alerts cover new findings and fixes that re-opened, not the same list again. Critical findings are loud, routine scans are quiet, and you choose which emails and webhook events you get.
How often do you re-check my app?
Turn on monitoring for a target and we re-check it every week and alert you when something changes. Monitoring is in early access. You can also re-run a scan any time, for example after a deploy; Pro includes 100 scans a month.
Do I need security expertise?
No. Each finding explains what is exposed and why it matters, with captured evidence and fix guidance you can use in Claude Code or Cursor. Review and test any suggested change before deploying it.
Can I use the results for SOC2, HIPAA, or a customer questionnaire?
You can export PDF reports with scan timestamps, and Team adds GDPR / SOC2 / HIPAA gap analysis and shared reports. If a customer or auditor needs a signed assessment, use a scoped pentest: it adds an agreed scope, authenticated multi-role testing, and a signed report.
Can I monitor several apps or client projects?
Yes. Pro covers up to 50 targets and 100 scans a month. Team adds 5 seats, scheduled scans with shared reports, and MCP and webhook API access.
How long does a scan take?
Initial triage arrives in under 60 seconds. A deeper automated scan typically takes 3–8 minutes, depending on scope and authentication flows. Human verification follows; its timing depends on the finding.

Need a scoped engagement with a signed report? Explore pentest services →

Try a tool. Read the research.

Find out what your app exposes.

Start with your live URL. Get evidence and a clear next step.

From $25 / month

Start free trial

14-day free trial · Cancel anytime

Not ready? Check one URL with the free Token Leak Checker, no signup.